usamike opened this issue on Mar 04, 2008 · 100 posts
usamike posted Tue, 04 March 2008 at 5:14 PM
Attached Link: http://renderfred.free.fr/p7benchmark.html
hi !Because i though the render of one single picture is very slow on my computer, i'm searching a way to benchmark Poser with other people hardware configuration. In the end, i'd like to buy a new computer, but i need advice and result of yours!
The best way would be someone create a standard ,quick and preinstall project like this one for Blender : http://www.eofw.org/bench/
Also, we could compare our render time for a same singular project.
But i'm wondering too, how can we display the rendering time process ? i didn't find this display option in Poser7 menus....Maybe come one could code a python script who display time before, and run render, and redisplay time...
Thanks for help for building a benchmark project.
So, by the way, i found this way to benchmark us :
Download this poser project file renderfred.free.fr/P7B.7z
It is compressed in 7-zip format, Winrar can unpack it.
Run Poser 7
3 Load/open the project file P7B.PZ3
4 Adjust number of threads (ctrl+K)
5 Hit render and start chrono (or open and look at window tray clock)
6 Stop/look at chrono (at render box exit).
7 report your result here please !
5m50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits
8m26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits
Thanks for play and participate !
usamike posted Tue, 04 March 2008 at 5:35 PM
i put a RAR version of the project here (17 Mb)
Then i just forgot to say that projet and benchmarl had been realized by http://renderfred.free.fr/
A new result (me again and for fun because my CPU have not 4 core, but i tried !!)
6m10s | 4 | Pentium D960 | 3600 | 2GB | XP 32 bits
usamike posted Wed, 05 March 2008 at 1:42 AM
no one wants to try ?
-blue- posted Wed, 05 March 2008 at 4:23 AM
Quote -
A new result (me again and for fun because my CPU have not 4 core, but i tried !!)Render TIME | Nb of Core | CPU | MHz | RAM | OS
6m10s | 4 | Pentium D960 | 3600 | 2GB | XP 32 bits
sorry to say but a pentium D960 does not have 4 cores and there are no dual socket 775 boards out there...
usamike posted Wed, 05 March 2008 at 4:31 AM
sorry, i meant 4 thread in poser, not 4 core !
did you make the test ?
please post your results here.
ghonma posted Wed, 05 March 2008 at 5:16 AM
I get about 3 mins on a q6600@2.4GHz, 4GB RAM, XP64 with 4 threads.
pjz99 posted Wed, 05 March 2008 at 5:30 AM
About 2 minutes 10 seconds for me.
Intel QX6700 overclocked to 3GHz
8GB RAM
XP64
I should add, this is not a great scene to benchmark with; over half the time from start to finish was spent on loading textures, something that for animation, would only need to be done once. Only one light in the scene is casting shadows, 512x512 shadowmap. I think a more serious benchmark would include reflection/refraction and AO to exaggerate the differences between machines.
usamike posted Wed, 05 March 2008 at 7:31 AM
thanks all, so now we have :
2m10s | 4 | QX6700 | 3000 | 8GB | XP 64 bits | pjz99
3m | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
5m50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
8m26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
let'sq go on !
noone else ?
tastiger posted Wed, 05 March 2008 at 11:22 AM
5m 40 s | 4 | AMD X2 5200+ | 2 x 2.61 GHZ | 4GB | Vista 64 bits
I also found about 1/2 the time was taken by loading textures
The supreme irony of life is that hardly anyone gets out of
it alive.
Robert A. Heinlein
11th Gen Intel(R) Core(TM) i9-11900K @ 3.50GHz 3.50 GHz
64.0 GB (63.9 GB usable)
Geforce RTX 3060 12 GB
Windows 11 Pro
thefunkyone_4ever posted Thu, 06 March 2008 at 12:33 AM
i get....
3m08s | 4 | E6600 | 2700 | 2GB | XP 32 bits
usamike posted Thu, 06 March 2008 at 12:59 AM
Now :
2m10s | 4 | QX6700 | 3000 | 8GB | XP 64 bits | pjz99
3m | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
3m08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunkyone
3m21s | 8 | 2x Opteron285 | 2600 | 8GB | XP 64 bits | renderfred
3m56s | 2 | E6400 | 2320ov | 2GB | XP 32 bits | renderfred
5m 40 s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
5m50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6m31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfred
8m26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
ghonma posted Thu, 06 March 2008 at 1:51 AM
How'd that guy set his threads to 8 ?
Anyway i did 2 more, one with the max overclock i can get and another with only 2 threads at stock.
2m07s | 4 | Q6600 | 3200 | 4GB | XP 64 bits | ghonma
4m02s | 2 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
Adom posted Thu, 06 March 2008 at 4:12 AM
2m15s | 4 | T7700 | 2400 | vista 64 | adom
2m25s | 3 | T7700 | 2400 | vista 64 | adom
2m25s | 2 | T7700 | 2400 | vista 64 | adom
4m00s | 1 | T7700 | 2400 | vista 64 | adom
nvidia quadro fx570m
After each render I closed Poser. Dont know why 2 & 3 gave same result.
Zanzo posted Thu, 06 March 2008 at 8:08 AM
Quote - About 2 minutes 10 seconds for me.
Intel QX6700 overclocked to 3GHz
8GB RAM
XP64I should add, this is not a great scene to benchmark with; over half the time from start to finish was spent on loading textures, something that for animation, would only need to be done once. Only one light in the scene is casting shadows, 512x512 shadowmap. I think a more serious benchmark would include reflection/refraction and AO to exaggerate the differences between machines.
Are you using the stock intel cpu fan that comes with the processor?
pjz99 posted Thu, 06 March 2008 at 8:25 AM
No, I have a Zalman CNPS9700:
http://www.newegg.com/Product/Product.aspx?Item=N82E16835118019
In addition to being an uber cooler, it is vewy vewy quiet, ideal for hunting wabbits. Typically the CPU fan is the biggest source of noise, and the included Intel cooling fan is noisy garbage.
Quote - How'd that guy set his threads to 8 ?
He didn't, OP inaccurately transcribed some of the information from Louguet's web page:
http://renderfred.free.fr/p7benchmark.html
MungoPark posted Thu, 06 March 2008 at 12:06 PM
Apple iMac Intel Core 2 Extreme at 2.8 ghz with 2gb memory 4 threads : 2.25
MungoPark posted Thu, 06 March 2008 at 12:10 PM
One addition: I used the internal render engine, because the external render engine is broken on Leopard 10.5
usamike posted Thu, 06 March 2008 at 3:58 PM
So, here we are :
2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
2"10s | 4 | QX6700 | 3000 | 8GB | XP 64 bits | pjz99
2"15s | 4 | T7700 | 2400 | ???? | vista 64 | adom
2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPark
3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunkyone
3"21s | 8 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfred
3"56s | 2 | E6400 | 2320ov | 2GB | XP 32 bits | renderfred
5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfred
8m26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
thanks again for continue ! don't hesitate to tell your friends !
pjz99 posted Thu, 06 March 2008 at 4:21 PM
You still have Louguet's benchmark transcribed wrong.
3"21s | 8 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfred
The Opteron is a 2-core processor, and at any rate Poser 7 can only work in 4 threads, not 8.
usamike posted Fri, 07 March 2008 at 1:34 AM
oups! thanks for the mistake noticed !
So, here we are :
2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
2"10s | 4 | QX6700 | 3000 | 8GB | XP 64 bits | pjz99
2"15s | 4 | T7700 | 2400 | ???? | vista 64 | adom
2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPark
3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunkyone
3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfred
3"56s | 2 | E6400 | 2320ov | 2GB | XP 32 bits | renderfred
5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfred
8m26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
People you know, i have the hight frequency cpu amongst this playlist (3.6 GHz), and i have the almost worst render time too !
This results are very meanfull for me ! i understand the L2 cache memory is more important than the frequency and (maybe) more important than the number of cores.
The little E6400 render more quickly than my D960 with less frequency, same core, but more L2cache memory.
Now i know, if i will have a Q6600 cpu, i will gain twice time in my works. Last august i bought my CPU 150€. Now the hightest cpu is, QX96500 and its cost is 830€. And i'm pretty sure with it, i won't gain 6x time (it is 6x more expansive than my cpu).
I really wonder what will be the results on XEON cpu...Does anybody know someone who could perform benchmark?
So, i really wonder how and when i could really perform Poser7 without waiting some long nights rendering...
You can see my first tries there http://www.dailymotion.com/mike_amiens_usa
All my video are full of little weaks but i 'm tired to re-render them : too long....
So...i hope another results will be come here!
ps : sorry for my expression syntaxe error, i 'm not english, even not american, just French.
ghonma posted Fri, 07 March 2008 at 2:07 AM
Quote - People you know, i have the hight frequency cpu amongst this playlist (3.6 GHz), and i have the almost worst render time too !
Well it's mainly cause the Pentium line is really old and Intel fixed most of it's mistakes with Core 2. Cache helps yes but the Core 2 was built from the ground up to be an AMD killer, which is also why the Q6600 is the best bang for the buck CPU you can buy (it was meant to kill AMD's Phenom) Xeons are great performers, but probably overkill for Poser.
Also you have to realize that it's not your CPU thats holding you back here, as much as firefly and poser's crappy way of handling textures. Like pjz99 said earlier, more then half the time in this scene is taken by poser loading textures, calculating shadow etc. The rendering itself takes like 30 sec on 4 threads.
To give you a comparison with what a good renderer can do, this also renders in under 3 mins (on mentalray):
That pic has a 100 apollos, all with SSS, GI, IBL, soft reflections etc. ie sometimes hardware is just not enough and you have to change your software as well.
MungoPark posted Fri, 07 March 2008 at 4:04 AM
I am encountering something strange - when I do the test render several times (always opening Poser new) it may be 2.25 and it may be 3.40 - Poser takes sometimes 90% of the cores and sometimes only 20% - there is no other process present in both cases - Mac users whats going on ?
magoo posted Fri, 07 March 2008 at 9:46 PM
2m 11s | 4 | QX6700 | 2670 MHz (stock) | 4GB | XP 64 bits
Had my comp running good @ 3.2 mhz with a 1600mhz FBS once, mebbe I'll try it 8-)
1m 59s on the very next render without closing poser down first.
xpac5896 posted Fri, 07 March 2008 at 11:12 PM
Just built this machine and wanted to know what it would do so here's the results
2m 46s | 3 | Q6600 | 2400 | 4 gig | XP 32
Here's my first test I did on the machine I just got rid of
7m 2s | 3 | P4+HT | 3200 | 2gig | XP 32
usamike posted Sat, 08 March 2008 at 2:19 AM
cool
2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
2"10s | 4 | QX6700 | 3000 | 8GB | XP 64 bits | pjz99
2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPark
2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunkyone
3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfred
3"56s | 2 | E6400 | 2320ov | 2GB | XP 32 bits | renderfred
5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfred
7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
thanks all again.
now i'm pretty sur which CPU i need now !
pjz99 posted Sat, 08 March 2008 at 8:42 AM
I hope you're not leaning towards the QX6700; currently it is the worst deal in terms of price/performance ratio.
Core 2 Quad Q6600 2.4GHz: $254US (best deal)
Core 2 Quad Q6700 2.66GHz: $540US (semi-bad deal)
Core 2 Extreme QX6700 2.66GHz: $985US **(worst deal!)
**Core 2 Extreme QX6850 3.0GHz: $980US (bad deal)
etc... The reason I bought a QX6700 was when I purchased, about a year ago, it was the only choice for a quad core processor; today, the Q6600 is a tremendously better purchase in terms of price/performance.
Blackhearted posted Sat, 08 March 2008 at 9:29 AM
IMO like pjz99 said, a true test of rendering speed should include something like a room with several basic objects in a cluster, all with AO, raytracing and using IBL, with production quality render settings.
most machines fly through simple renders, and a 1-2 minute difference is not as significant here. if so much time is being wasted just loading textures, simple factors like drive read speed or fragmentation could have more significant an impact on the render than the actual processing power.
with a hefty scene youll start seeing overheating, pagefile usage, out of memory errors, etc. this would be, IMO, a better indicator for people deciding on what computer to buy. judging by the machines and rendertimes i could probably render this scene on a 500mhz P3. you need a scene that brings more performance factors into play.
usamike posted Sat, 08 March 2008 at 12:03 PM
Blackhearted, i'm pretty sure you're right.
i just felt on this projet, with some already result. so i forward it here.
at least, it's a samll kind of reference for beginner project in poser.
i'd be glad if someone (what about you) could build a projet with standart objects of Poser 7 sr2.1 (no addon, no daz3d addon, .no renderosity addon), just what is in the poser7 package.
magoo posted Sat, 08 March 2008 at 3:06 PM
Like pjz99, I built my machine 10 months ago and the only quad available was the QX6700. But
if I were going to built today, I'd go for the QX9650. Only because it's a 45nm architecture chip,
and the extreme chips are unlocked (better for overclocking). Now if money was an issue, I'd
wait for the regular quads to come out with the 45nm, these 65nm chips run hot, way HOT. But
all in all, I'm very happy with what I have now. Haven't been able to bring it to it's knees yet. And I
love loading up scenes. 8-)
usamike posted Sat, 08 March 2008 at 4:34 PM
hey maggo ! shall you explain me what is the green thing in upper picture ?
it looks like a.....hum....how can i mean..well...an alien with a human dick shape ! moreover it's got a frog skin !
magoo posted Sat, 08 March 2008 at 5:50 PM
you hit the nail on the head, it's an alien. 8-)
magoo posted Sat, 08 March 2008 at 6:51 PM
1m 57s | 4 | QX6700 | @3000 with 1333FSB | 4GB | XP 64 bits
8-)
didn't hit render a second time, probably would've been like 1m 45s one would think.
usamike posted Sun, 09 March 2008 at 3:42 AM
1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPark
2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunkyone
3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfred
3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfred
5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfred
7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
urbanarmitage posted Wed, 19 March 2008 at 5:45 AM
I've been meaning to try this on my desktop and notebook for a while. Here are the results for my notebook (Dell D830) and i'll do my desktop in the next couple of days too. The results so far look interesting ...
4'43" - 2 threads (one HT) - Mobile Core 2 Duo T7700 - 2400Mhz - 4 Gig RAM - XP Pro 32-bit
6'03" - 1 thread - Mobile Core 2 Duo T7700 - 2400 Mhz - 4 Gig RAM XP Pro 32-bit
UA
usamike posted Sat, 22 March 2008 at 10:44 AM
hi again people,
i'm projecting to change my computer. and i think about one point :
With POSER7, Is a rendering operation speeder under 64bits than under 32bits (bot with 4BM and core2Quad CPU) ?
Then, in this case, is this faster under XP64 or Vista64 (same hardware and poser7) ?
any idea ?
pjz99 posted Sat, 22 March 2008 at 11:39 AM
64-bit operating system has little effect on render speed (possibly can make it a very small bit slower, on the order of 1-3% from what I remember of Louguet's info). Since Poser is currently only available in 32-bit versions, this is not a big benefit, but I've found it does improve memory usage problems. On the other hand if you use other apps that ARE written in 64-bit instructions, you basically never have an app crash because it ran out of memory. I have 8GB of memory, and I remember one instance in Cinema4D (64-bit) where I converted a complicated dynamic hair setup to polygons - many, many millions of polygons. And then I rendered it. The process took up 20gb of memory (RAM + virtual), but it did not crash. A couple of times Windows complained that it needed to enlarge the swap file, but the render just went right on and eventually completed normally. I was very impressed.
I use Windows XP Pro 64-bit edition and recommend it. Some people are very happy with Vista 64, but while I have heard many complaints of application compatibility under Vista 64 I have heard virtually none about XP Pro 64. All of my apps work VERY well.
I strongly recommend that you do not cut any corners on processor, motherboard, or video card, all three items are heavily used in 3D regardless of application. I think an 8xxx series Nvidia card is the safest and fastest option for video. Good luck!
svdl posted Sat, 22 March 2008 at 12:46 PM
There's a little caveat with XP Pro 64 bit. Driver support. I recently build another Q6600 system, and no matter what I tried, I couldn't get the mainboard at full functionality under XP Pro 64 bit, and Poser proved to be unstable in the XP 64 setup.
Switched to Vista 64 bit, and now the mainboard is fully functional and Poser is stable (in general).
My other Q6600 setup, which is a year older, runs stable on XP Pro 64 bit.
Funny thing is that there is quite a speed difference between the two systems, using the same CPU, not overclocked. The newer system is significantly faster.
It might have to do with the RAM (the older system runs at 667, the newer at 800), or with the chipset (both systems have a P965 chipset, but different southbridge chips). The disks in the "old" system should be faster than the disks in the "new" system.
I'm not taking OpenGL performance into account. The "old" system has a 7800GTX graphics card - no slouch, but not as fast as the 8800GT in the new system.
The pen is mightier than the sword. But if you literally want to have some impact, use a typewriter
pjz99 posted Sat, 22 March 2008 at 1:58 PM
BIOS and chipset have a huge impact on overall speed. Not knowing which chipsets or BIOS you have it's hard to say but you may want to investigate that. I can recall some cases where people updated their BIOS and saw a noticeable performance bump. There is also stuff like the PCI Express speed problems as Gareee posted about in this thread:
http://www.renderosity.com/mod/forumpro/showthread.php?thread_id=2734605
Chippsyann posted Sat, 22 March 2008 at 2:37 PM
2min 23sec | 2 | Quad Core | 2.40 | 8GB | XP 64 bits
Q6600
1min 58sec | 4 | Quad Core | 2.40 | 8GB | XP 64 bits
Q6600
Here's my numbers, for your test.
Penguinisto posted Sun, 23 March 2008 at 12:26 PM
Quote - > Quote - People you know, i have the hight frequency cpu amongst this playlist (3.6 GHz), and i have the almost worst render time too !
Well it's mainly cause the Pentium line is really old and Intel fixed most of it's mistakes with Core 2. Cache helps yes but the Core 2 was built from the ground up to be an AMD killer, which is also why the Q6600 is the best bang for the buck CPU you can buy (it was meant to kill AMD's Phenom) Xeons are great performers, but probably overkill for Poser.
No Poser 7 here (oops), but I do agree w/ the above, and here's what happened: The Early Pentium series relied on a (IMHO crap) hardware architecture called NetBurst. Long/complex story short, Netburst introduced a shedload of internal latency in the CPU itself due to an excess of required stages that data had to sit through in there, which they tried to compensate against with insanely high clock speeds (they originally wanted to push things to 10GHz, but hit a brick wall w/ 4). Problem is, there still was enough lag that showed, especially against AMD chips coming out against it at the time. Intel decided they couldn't keep doing it w/ Netburst, so the went back to the philosophies of the late Pentium III line, and simplified things greatly with the Pentium M series (originally just their P4 laptop architecture - the "M" meaning "Mobile"). Pentium M put more emphasis on data-volume-per-cycle and in energy efficiency. The results of that are seen in the Core (and Viiv) series. Now at the time, AMD's Opteron was eating Intel's lunch. But Intel did do one thing that most corps are not normally willing to do - they stepped back, admitted their mistakes, and took a holistic approach towards not repeating them. When the Core line came out, it was pretty revolutionary. The emphasis was no longer on pumping up clock speeds, but a combination of two things: efficiency (in both power and data movement), and physically shrinking the distance that data has to travel. For instance, a 45nm process chip (45nm = 45 nanometer min. spacing between electrical traces on a chip) is half the size of a 90nm chip of the exact same design, so the data only has to travel (roughly) half as far in total, giving you some pretty amazing performance increases. Notice that it doesn't double the performance (for many reasons), but it contributes hugely towards increasing the speed at which a given instruction can be executed. The biggest reason Intel has been able to eat AMD's lunch recently has to do with the tech behind this shrinking tech. AMD is just now barely getting it on with 45nm chip fabrication (with troubles, as evidenced by the troubles in the Phenom line - the Opteron is 90nm still). Intel meanwhile has enjoyed selling 45nm tech for well over 18 months or so, and is just about to come out with 35nm tech. Of course, you can only shrink something so much, until electron leakage and existing fab tech hits a brick wall (let's face it 35 nanometers is barely enough room for some complex molecules). This is why multi-core application is getting so popular. And so it goes... :) HTH a little. /P (Disclaimer: I work in and around the semiconductor industry).
urbanarmitage posted Sun, 23 March 2008 at 3:22 PM
Ok, well here are the results from my desktop machine. They are quite a bit better than my notebook which isn't all that surprising even though my desktop has an older processor. I did find Poser very unstable when loading the test render PZ3 even though generally I have found it to be quite stable under XP 64-bit. I had to restart my desktop a few times because Poser would crash just afterloading the PZ3.
The most interesting thing was the fact that even though my desktop has a dual-core CPU, bumping the number of threads up to 4 improved the render time by 1 second. I have noticed that sometimes using 4 threads takes a bit longer to render but in this instance it saved me a whopping 1 second! :biggrin:
To keep things in perspective, my desktop has a SATA2 Seagate Baracuda hard drive running at 3GB/s bandwidth and my notebook still has a PATA hard drive.
04'41" - 1 thread - AMD Athlon 64 X2 Dual-core 6400+ Black Ed - 3200Mhz - 4Gig DDR2-800 - XP Pro 64-bit
03'07" - 2 threads - AMD Athlon 64 X2 Dual-core 6400+ Black Ed - 3200Mhz - 4Gig DDR2-800 - XP Pro 64-bit
03'06" - 4 threads - AMD Athlon 64 X2 Dual-core 6400+ Black Ed - 3200Mhz - 4Gig DDR2-800 - XP Pro 64-bit
UA
usamike posted Mon, 24 March 2008 at 11:58 AM
1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAnn
2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAnn
2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPark
2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunkyone
3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfred
3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfred
4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanArmitage
5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanArmitage
6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfred
7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
thanks all again !
svdl posted Mon, 24 March 2008 at 3:53 PM
Running with 4 cores enabled:
4 threads: 1"13s
3 threads: 1"27s
2 threads: 1"51s
1 thread: 2"55s
Running with only 2 cores enabled:
4 threads: 1"39s
3 threads: 1"43s
2 threads: 1"50s
1 thread: 2"55s
Timed using the attached TimedRender python script, adapted from gwhicks TimedRenderCache script.
Remarkable results. The Q6600 is not the fastest CPU in the test, still the test results are the quickest by far. I suspect that the fact that Poser is installed on its own physical drive helps, and it confirms what has already been said: this test depends heavily on other factors than CPU, mainly on disk related factors.
The pen is mightier than the sword. But if you literally want to have some impact, use a typewriter
pjz99 posted Mon, 24 March 2008 at 4:02 PM
Much of the reason for smaller difference is that this particular benchmark is not a great overall test - a great deal of the render job with this scene is preparing it for the renderer itself, loading objects and texture data, and it's all done in one thread (a great pity). Although it is good info, a more thorough benchmark would have things like reflection, AO, and the like - although Poser is still not very good at the multitasking business as I think we've discussed before, because render threads that finish early leave you with idle cores, and the work is not re-divided to make the idle cores busy again. Depending heavily on how the scene is laid out, e.g. if you have tons of reflection in the upper right corner of the frame, having multiple cores may not be a very big benefit there.
urbanarmitage posted Tue, 25 March 2008 at 2:36 AM
SVDL, very nice render times there! I had a Q6600 in my hands at the PC shop the other day, turning it over and over muttering something like 'my precioussssss!!!!!'. :biggrin:
The Q6600 seems to perform very well in this comparison. I agree though that the initial loading of textures and preparing for the render takes up a large chunk of the render times. On my AMD it was about 1'45" into the render that both cores kicked in and on my notebook it was closer to 3'00". The hard drive throughput is having a huge impact on that, as is memory timings IMHO.
UA
gsfcreator posted Tue, 01 April 2008 at 9:15 AM
i tried twice with my pc. CPU=Q6600, 4GB RAM(but only 3 recognizable by my xp32bit OS). mhz=2.4
first trial - 4threads = 3:40 seconds(2 minutes + for texture loading:|! )
second trial - 4threads, seperate process - around 2:32 seconds(but maybe its cause its the second trail? i havent closed poser between thr trials :)
dalroi27 posted Thu, 03 April 2008 at 3:26 AM
2"11 | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
usamike posted Mon, 14 April 2008 at 1:59 AM
I hope more mobile CPU like T7700 will make the render test because i think about purchase one powerfull laptop only for render process (also i need a loptop too).
So i should go on my work on Desktop while rendering on the laptop.
Actually i think about X9000 or T9500 mobile CPU config , 4GB, OS64 and with fast HD and GeForce™ Go 8600M GT - 256 (alienware or Dell)...if anybody have one and poser on his laptop...please make the test !
update the list !
1"13s | 4 | Q6600 | 2400 | 8GB | Vista64 bits | svdl
1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAnn
2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
2"11s | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAnn
2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPark
2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunkyone
3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfred
3"40s | 4 | Q6600 | 2400 | 4GB | XP 32 bits | gsfcreator
3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfred
4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanArmitage
5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanArmitage
6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfred
7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
thanks all again !
softcris posted Mon, 14 April 2008 at 6:27 PM
well a bit disapointed with my bench mark...since I got a Q6600. a SLI 8800GTS (512 MB each one) 8 GB memory OCZ extreme- running XP 64 bits just clean from formatting about 2 hours ago (changed the board to a more suitable- 780i XFX >TriSliWay)
that same scene runs the render at it's..**for about 4 min.
Also tested the Cloth room with a V4 and Chinadress with 90 frames - took 13 min. Render it 4 min. What is wrong here? got a good hardware but results are not that good!
2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
2"10s | 4 | QX6700 | 3000 | 8GB | XP 64 bits | **pjz99
**2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | **MungoPark
**2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | **xpac5896
**3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | **ghonma
**3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | **thefunkyone
**3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | **renderfred
**3"56s | 2 | E6400 | 2320ov | 2GB | XP 32 bits | **renderfred
**5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | **tastiger
**5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | **renderfred
**7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | **xpac5896
**8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
4" | 4 | Q6600 | 2400 | 8GB | XP 64bits | softcris
"'you shut up! or I'll
bring democracy to your country! "
Cris
Galvão aka Softcris - www.crisgalvao.com
(or softcris,
SoftCris)
Rendering since 1997 and
at Renderosity since 1999.
OS
Win 8.1 64 bit
usamike posted Tue, 15 April 2008 at 12:51 AM
hi softcris !
yes i saw many difference between Q6600 result (even with no overclock).
pjz99 posted Tue, 15 April 2008 at 1:48 AM
Are you certain you have Poser 7 set to use 4 threads? If so then I'd suspect your chipset, it shouldn't be THAT slow. I'm still not sold on SLI automagically speeding up OpenGL either, I've heard from far too many authoritative sources that it provides no benefit unless the app is specifically written for SLI, or can actually slow you down.
Note also, the Cloth Room is single threaded no matter what, so having multiple cores won't benefit you there.
usamike posted Tue, 15 April 2008 at 5:17 AM
Quote -
Note also, the Cloth Room is single threaded no matter what, so having multiple cores won't benefit you there.
that's why i'm proud of my old architecture and high natural frequency cpu Pentium D960@3.6GHZ !
Because there is many many tasks to do on a desktop computer that are not multi-core/thread yet !
magoo posted Tue, 15 April 2008 at 8:12 AM
softcris there was a new bios released at the end of March for the 780i, and new mother board
drivers are due out soon, if not already. Check you bios version on post, then go to the XFX site
and see what the latest is, and look for the MB drivers.
8-)
softcris posted Tue, 15 April 2008 at 1:49 PM
hello there!
yesterday when I posted here was already over mid night here...got out early to work so....
yes, guys; I have being since 6 PM till mid night w/ the technician installing the '780i' XFX in the machine.
Poser 7 sp 2. Sli enable.
All good but not higher speed.
Thanks for the support. I continue this in Hardware post under: http://www.renderosity.com/mod/forumpro/showthread.php?thread_id=2735431&page=2#message_3210408
"'you shut up! or I'll
bring democracy to your country! "
Cris
Galvão aka Softcris - www.crisgalvao.com
(or softcris,
SoftCris)
Rendering since 1997 and
at Renderosity since 1999.
OS
Win 8.1 64 bit
flaviok posted Mon, 14 July 2008 at 7:37 AM
1"30scs
thread 4 ( separate process)
AMD ATHLOM 64 X2 6.600
4 GB RAM DUAL 800
VISTA 64
ATI RADEON HD 2600
usamike posted Mon, 09 March 2009 at 4:19 AM
shall somebody with a core i7 CPU make the test please ?
12rounds posted Mon, 09 March 2009 at 11:03 AM
Processor: i7 920 2,6Ghz (quad core)
Motherboard: P6T Deluxe
OS: Ubuntu Linux 64-bit
Best time: 75 seconds (4 threads)
I agree with the criticism about the benchmark; most of the time goes to non-threaded activities such as loading textures and rendering shadow maps and such.
12rounds posted Mon, 09 March 2009 at 1:59 PM
I improved the above figure of 75 seconds to 66 seconds by two simple tricks.
I first observed that Poser wants to write every texture it reads before it switches to the rendering engine into it's "PoserTextureCache" folder. In this case the amount of data written is around 1,1Gb (Poser writes ALL the texture data to .exr file format probably because FireFly wants it that way). Poser also behaves very stupidly in this ... if it encounters multiple instances of the same texture file (like in this case "KateShoes" and "KateShirt" etc, it just writes the file again ... and again... and again. That's why you may see some texture files disappear quickly from the "Loading textures" phase and some seem to stay there forever.
Okay. Writing 1,1Gb creates a small amount of i/o traffic. Not enough to choke i/o operations (at least on my system) since the actual process that writes the files is a single thread and even if the disks and busses could take more load, Poser just isn't pushing them to the limit. Nevertheless taking that into notice, I shaved a few seconds by pointing the texture cache to a RAM disk (actually a "tmpfs" filesystem in Linux, but I suppose it's a RAM disk in Windows). That pretty much eliminates any i/o overhead caused by writing the said 1,1Gb.
Next in the general preferences tab the default values for "Memory limit buffer" and "Adaption threshold" are not optimised for my system (I did a fresh P7 install just for this benchmark since I had already deleted most of the needed resources required by the pz3). So I went to the upper limit and put 2047 as the "Memory limit buffer" and "0" as the adaption threshold. With these two things ... over 10% total speed increase.
Normally I wouldn't want to use RAM as the renderer's texture cache because of the obscene file sizes of .exr images.
stewer posted Mon, 09 March 2009 at 2:20 PM
Quote - Poser also behaves very stupidly in this ... if it encounters multiple instances of the same texture file (like in this case "KateShoes" and "KateShirt" etc, it just writes the file again ... and again... and again.
Are those files dated in the future by any chance?
For managing the texture cache, Poser checks the file dates of the source textures and compares them to the files dates mip maps in the cache. This way it can react appropriately when files get changed in Photoshop between renders.
12rounds posted Mon, 09 March 2009 at 3:07 PM
Quote - > Quote - Poser also behaves very stupidly in this ... if it encounters multiple instances of the same texture file (like in this case "KateShoes" and "KateShirt" etc, it just writes the file again ... and again... and again.
Are those files dated in the future by any chance?
For managing the texture cache, Poser checks the file dates of the source textures and compares them to the files dates mip maps in the cache. This way it can react appropriately when files get changed in Photoshop between renders.
That's very interesting.
The original files (Kate resource files) referenced by the pz3 are dated 13 of Nov. 2006 on my system so I can't see why that would be happening. Cause for further investigation as it really slows down this benchmark for me (clicking open some of my own pz3 files doesn't make this happen so it's an anomaly and not a norm).
usamike posted Mon, 09 March 2009 at 5:25 PM
thanks for all your remarks
i knonw that benchmark is not a reference, but it's the one ! And somewhere, it provides any idead of gain by CPU to another; at least a minimal expected performance if nobody did tweak his hardware before the test !
You know i own a pentium D960, i will purchase a core i7 (maybe i7 940 or 950). So by reading thi all your score, i guess i will gain x4.5 speed at leastn wheras in cinebench i will gain only x3, and in Blender x5 (http://www.eofw.org/bench/)
In my drawer, i have a compactflash-to-SATA adaptator and a 266x compact flash card. I forecast to build a speed SSD with this (for only 100€) to use for a system cache partition (where poser will be)
thanks again for all !
ghonma posted Mon, 09 March 2009 at 9:40 PM
Quote - In my drawer, i have a compactflash-to-SATA adaptator and a 266x compact flash card. I forecast to build a speed SSD with this (for only 100€) to use for a system cache partition (where poser will be)
thanks again for all !
Flash cards are actually much slower then modern hard disks. Any recent HDD of 1+ TB with decent amount of cache can do a burst transfer of over 100MB/s and sustained of over 70MB/s Compared to this flash cards, even the 266x can do only 10-15MB/s max.
IMO you'd be much better off getting a WD caviar 1TB or something like that, which will cost you about the same 100€ and will give you a lot more space to play with.
grichter posted Mon, 09 March 2009 at 10:49 PM
What can throw the results off is if you render twice without quitting Poser as it appears you now have the textures in cache
MacOSX 10.5.6, Quad Core Intel Xeon 3 ghz, 16 gigs of 667DDR2 Ram, 8 cores, PoserPro, 4 threads, render as a separate process checked.
First render in background 2:03
Second Render Internal 1:01 (did not quit between renders)
Third Render again in Background 0:58 (did not quit between renders)
Quit and restart Poser
Fourth Render Internal 2:21
Fifth Render again in Background 0:58 (did not quit between renders)
After a fresh launch of Poser, it takes about half the Render time just loading the textures.
Gary
"Those who lose themselves in a passion lose less than those who lose their passion"
12rounds posted Tue, 10 March 2009 at 1:33 AM
Quote -
In my drawer, i have a compactflash-to-SATA adaptator and a 266x compact flash card. I forecast to build a speed SSD with this (for only 100€) to use for a system cache partition (where poser will be)
I would advise against it. Not only flash cards are slow to write to, they have a limited amount of write cycles per memory cell and have no wear-levelling, no write buffers, no read caches etc.
Then again if you're going for a real SSD, THAT I would say will get you get nice big boost - for a pricetag. My experiments with a OCZ 60Gb SSD housing Poser are showing a huge impact on loading times. But I also put caches to a 10000rpm disk since SSD writes don't give the same benefits for writing as they do for reading and also this way also the system doesn't have to read and write the same disk at the same time. While my WD Velociraptor is THE fastest HD I've yet to encounter (for home use, that is), it loses to my SSD in every possible measurement I've made (the most important being a real-time analysis of disk activity during actual Poser use).
usamike posted Tue, 10 March 2009 at 6:10 AM
well, i guess i need this solution :
12rounds posted Tue, 10 March 2009 at 6:41 AM
Quote - well, i guess i need this solution :
LoL!
In that video they did a defrag on the SSDs... that's just stupid.
Defragmentation is pretty much meaningless when SSDs are concerned and defragging the disk can be even downright harmful - certainly it doesn't give any benefits.
usamike posted Sat, 02 January 2010 at 10:10 AM
i update the records :
update the list !
1"13s | 4 | Q6600 | 2400 | 8GB | Vista64 bits | svdl
1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAnn
2"01s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
2"11s | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAnn
2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPark
2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunkyone
3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfred
3"40s | 4 | Q6600 | 2400 | 4GB | XP 32 bits | gsfcreator
3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfred
4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArmitage
4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanArmitage
5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanArmitage
6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfred
7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
well ! i you can see ! it is plenty not a good test , neither a good benchmark !
i didn't say that because i have a huge hardware (red line) and i am not be on the top of the list.
Everybody said that !
It would be good to have a real benchmark for poser like cinebench10 is for c4d !
A benchmark which can handle 8 threads and compiled in 64bits natively (will it be poser pro 2010 ?)
well , let's see now with poser 8, i'm installing it on the same hardware.
usamike posted Sat, 02 January 2010 at 11:04 AM
i updated again the records :
what did it change ?
i add new columns for poser version (V), and for the number of the successive render (T).
As the scene file can run in both poser 7 and 8 (maybe poser pro ?), it is very usefull to compare.
As the first render (directly after the scene loading) is longer than the second (due to hard drive read and caching), it is important to know when the render had been done.
in my case (usamike), the first loading of the scene takes 1min before render and light computing. And in the second render, the hard drive didn't run (led is off) so i gess all the texture are been loaded from the memory cache (6 MB is enough to hangle the complete scene and its textures).
That's why in the previous result, the time render vas very curious (very longer than a q6600)
And known, on the same machine i have poser 7 and poser8 trial, and i can have an idea of the factor of render improvement :
seoncd render :
p7 : 47s
p8 : 33s
first render : (same time of hard drive loading, it is stranged..)
p7 : 2"01s
p8 : 1"33s
so at last, with 4 more threads (p8 uses my 8 threads where p7 uses 4), poser 8 renderfirefly is x1.42 quicker than poser 7.
i post there the render image in poser7 and poser8 to be sure it is the same in my eyes.
i looked it very closer, i didn't see any obvious differences except the size of the files.
poser7
poser8
if someone have a picture tool to compare it pixel by pixel, it can be usefull but it is enought for my eyes !
**8 | 2 | 0"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike **
7 | 2 | 0"47s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 1"13s | 4 | Q6600 | 2400 | 8GB | Vista64 bits | svdl
8 | 1 | 1"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
7 | ? | 1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAn
7 | 1 | 2"01s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
7 | ? | 2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
7 | ? | 2"11s | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
7 | ? | 2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
7 | ? | 2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
7 | ? | 2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits |ChippsyAnn
7 | ? | 2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPa
7 | ? | 2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
7 | ? | 3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
7 | ? | 3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunky
7 | ? | 3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfre
7 | ? | 3"40s | 4 | Q6600 | 2400 | 4GB | XP 32 bits | gsfcreato
7 | ? | 3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfre
7 | ? | 4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
7 | ? | 5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
7 | ? | 5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
7 | ? | 6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
7 | ? | 6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfr
7 | ? | 7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
7 | ? | 8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
usamike posted Tue, 05 January 2010 at 9:17 AM
updated by ricky1:
8 | 2 | 0"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | 2 | 0"47s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 1"13s | 4 | Q6600 | 2400 | 8GB | Vista64 bits | svdl
8 | 1 | 1"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
7 | ? | 1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAn
7 | 1 | 2"01s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 2 | 02:00 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
7 | ? | 2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
7 | ? | 2"11s | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
7 | ? | 2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
7 | ? | 2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
7 | ? | 2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits |ChippsyAnn
7 | ? | 2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPa
7 | ? | 2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
7 | ? | 3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
7 | ? | 3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunky
7 | ? | 3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfre
7 | ? | 3"40s | 4 | Q6600 | 2400 | 4GB | XP 32 bits | gsfcreato
8 | 1 | 03:45 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfre
7 | ? | 4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
7 | ? | 5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
7 | ? | 5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
7 | ? | 6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
7 | ? | 6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfr
7 | ? | 7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
7 | ? | 8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
Michael314 posted Tue, 20 April 2010 at 11:57 AM
Hello,
just stumbled over this thread. Here's some data from my PC. I don't have Poser 7 installed any more, so I used P for PoserPro, and 2 for PoserPro 2010.
The data shows how much better the renderer now is, apart from that my system seems to be much slower than comparable boxes. I use standard memory timings, maybe that's the problem.
2 | 2 | 0"37s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
8 | 2 | 2"06s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
P | 2 | 5"20s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
2 | 1 | 1"55s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
8 | 1 | 2"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
P | 1 | 5"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
Best regards,
Michael
Michael_REMY posted Tue, 20 April 2010 at 4:47 PM
thanks you !
i have poser pro 2010 too, i will make the test next week too.
here now, the list :
8 | 2 | 0"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 2 | 0"37s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | 2 | 0"47s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 1"13s | 4 | Q6600 | 2400 | 8GB | Vista64 bits | svdl
8 | 1 | 1"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 1 | 1"55s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
7 | ? | 1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAn
7 | 1 | 2"01s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 2 | 02:00 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
8 | 2 | 2"06s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
7 | ? | 2"11s | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
7 | ? | 2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
7 | ? | 2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
7 | ? | 2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits |ChippsyAnn
7 | ? | 2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPa
7 | ? | 2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
8 | 1 | 2"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
7 | ? | 3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunky
7 | ? | 3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfre
7 | ? | 3"40s | 4 | Q6600 | 2400 | 4GB | XP 32 bits | gsfcreato
8 | 1 | 03:45 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfre
7 | ? | 4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
P | 2 | 5"20s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
P | 1 | 5"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
7 | ? | 5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
7 | ? | 6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
7 | ? | 6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfr
7 | ? | 7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
7 | ? | 8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
LostAlien posted Sun, 09 October 2011 at 6:02 AM
Hello, Just ran the test for you...
Hope that helps
LA
usamike posted Sun, 09 October 2011 at 2:12 PM
good Job lost alien ! i will do the same in the week (i just acquire&receive my pp2012) on my same and current workstation (ci7-975).
Your result is stomached me ! nowhere in my dream, i would have exptected a such improvment for a 3d software !
LostAlien posted Tue, 11 October 2011 at 4:14 PM
usamike,
Glad you are happy with the result.
I do have a bit of a monster PC as well... It all helps!
Let me know if you want me to test anything else.
LA :biggrin:
Michael_REMY posted Sun, 26 February 2012 at 5:11 AM
Quote - Hello, Just ran the test for you...
- Poser Version : 2012 pro 64bit
- Render time : 0 minutes and 20 seconds
- CPU : i7-2820QM
- Threads : 8
- MHz : 2.30
- RAM : 8GB + 3GB Video
- OS : Windows 7 Ultimate
- User : Lost Alien
Hope that helps
LA
you score is wrong because poser2012 miss several texture of the g2 model.
you might had error message while loading p7b file.
i miss this :
G2Casual-TennisShoes
G2Casual-T-shirt
G2Casual-CasualPants
G2_Pants
G2_T_shirt
G2_Tennis_Shoes
so, 2 models (of 4) are nude, without clothes textures. Then the render is obviously faster !
i have to find out where these texture are gone ?
aRtBee posted Sun, 26 February 2012 at 7:15 AM
Glad to give some info too. Scene downloaded, unpacked and run, no flaws.
Score: 14-15 seconds to render, either internal or as external process (consistent over 6 runs).
When rendering external, shadow maps have to be calculated first unless re-used from an earlier run (check the menu option), this adds 5 to 10 sec and seems to be a 1-thread process. When rendering internally, things seem to be slightly different (= much faster).
When rendering the first time textures have to be found, this takes 40-50 sec additionally. Its a 1-thread process whether you render internally or externally. Speed depends on the size of the Poser Runtime folders which have to be searched, mine are 45-50Gb, the default Poser one is about 10% of this. The search result is stored in a Temp folder, so the search has to be redone only when this is cleared between scenes or Poser launches or - as in my case - machine reboots.
Specs:
win7 Pro Sp1 64-bit, PPro 2010 64-bit, CPU: i990X @4GHz (6 cores, 12 threads), RAM: 24Gb incl 4Gb Ram-disk for Temp, C: on SSD (program only, no data), Data on 9x2Tb disk unit, nVidia GTX560Ti OC/1Gb.
Indirect Lighting anyone? I just checked the box in Render Settings.
4.5 min (=270 sec) preparation, 95 sec rendering = 365 total. That's say 25 times as long compared to No IDL. No additional memory required. CPU temp ramped up to 75C = 170F (Idle=45*C). This is fun. Note that using LuxRender might need 30 mins to get a full blown photoreal result (well, not for this scene I guess), which is again 5 times longer.
- - - - -
Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.
visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though
Michael_REMY posted Sun, 26 February 2012 at 9:37 AM
Quote - Glad to give some info too. Scene downloaded, unpacked and run, no flaws. Score: 14-15 seconds to render, either internal or as external process (consistent over 6 runs).(...)
Specs:
win7 Pro Sp1 64-bit, PPro 2010 64-bit, CPU: i990X @4GHz (6 cores, 12 threads), RAM: 24Gb incl 4Gb Ram-disk for Temp, C: on SSD (program only, no data), Data on 9x2Tb disk unit, nVidia GTX560Ti OC/1Gb.
so, according to your specs, i guess you use poser pro 2010 (not 2012) to run the test. that's explan why you did not have the miss texture file problem.
And for the record, with your i990X @4GHz and 12 threads and SSD configuration, you reached a x2 factor faster than my cpu (i7-975 3.3Ghz and 8 thread and hard drive). What is pretty much not bad at all !
aRtBee posted Sun, 26 February 2012 at 10:23 AM
yeah, PPro2010 was mentioned in the specs.
And indeed: (12 threads * 4GHz) / (8 threads * 3.3Ghz) = 180%. Disks don't do much in this scene, and if any: Ram-disk beats all. Its simply a matter of horsepower. I guess UsaMike is aware of this (by now).
In reverse: my 14 sec * 180% = 25.2 which is 5 sec above your 20, hence PPro2012 might be 20% faster when taking the HW-factor out. Nice to know. I'm curious what they did to the IDL handling, perhaps you can give it a try. Total prep+render should take my 365 sec * 180% HW-factor * 80% PPro2012-factor = 525 sec.
- - - - -
Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.
visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though
usamike posted Sun, 26 February 2012 at 12:01 PM
updated on 26 february 2012
For whom wants to redo the test, i remember links (with textures fixed):
for poser pro 2012 only users : http://www.michael-remy.fr/forum/pro2012.zip
for poser pro 2010 users : http://www.michael-remy.fr/forum/pro2010.zip
for other poser (7,8 maybe 6) users : http://www.michael-remy.fr/forum/poser7.zip
legend :
v (version) :
2 : poser pro 2012
0 : poser pro 2010
8 : poser 8
7 : poser 7
p : poser pro (release before the 2010 one)
0 | ? | 0"14s | 12 | Ci7-990X | 4000 | 24GB | W7 64 bits | Artbee (poser p2010 and SSD)
2 | ? | 0"20s | 8 | Ci7-2820QM | 2300 | 8GB | W7 64 bits | Lost Alien (maybe missing texture)
2 | 2 | 0"22s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 1 | 1"11s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
0 | 2 | 0"25s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
0 | 1 | 1"15s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 2 | 0"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 2 | 0"37s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | 2 | 0"47s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 1"13s | 4 | Q6600 | 2400 | 8GB | Vista64 bits | svdl
8 | 1 | 1"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 1 | 1"55s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
7 | ? | 1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAn
7 | 1 | 2"01s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 2 | 02:00 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
8 | 2 | 2"06s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
7 | ? | 2"11s | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
7 | ? | 2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
7 | ? | 2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
7 | ? | 2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits |ChippsyAnn
7 | ? | 2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPa
7 | ? | 2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
8 | 1 | 2"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
7 | ? | 3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunky
7 | ? | 3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfre
7 | ? | 3"40s | 4 | Q6600 | 2400 | 4GB | XP 32 bits | gsfcreato
8 | 1 | 03:45 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfre
7 | ? | 4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
P | 2 | 5"20s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
P | 1 | 5"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
7 | ? | 5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
7 | ? | 6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
7 | ? | 6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfr
7 | ? | 7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
7 | ? | 8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
WandW posted Sun, 26 February 2012 at 1:05 PM
AMD Phenom X6, 3200 MHz, 16 GB, 6 threads Poser Pro 2012; 1' 37"
For this to be a true CPU benchmark, rather than a disk benchmark, one should hit Cancel after it starts rendering, so the textures remain loaded in RAM, and then rerender. Doing that, I got 38 seconds. Going from a 64 pixel to 128 pixel bucket size cut off 0.3 seconds... 
EDIT Switching light 2 to Raytraced shadows cut it down to 28 seconds.
----------------------------------------------------------------------------------------
The Wisdom of bagginsbill:
"Oh - the manual says that? I have never read the manual - this must be why."
WandW posted Sun, 26 February 2012 at 1:27 PM
Oops; too late to edit.
I inadvertently turned shadows off above 
It took 1' 2" with raytraced shadows...
----------------------------------------------------------------------------------------
The Wisdom of bagginsbill:
"Oh - the manual says that? I have never read the manual - this must be why."
aRtBee posted Sun, 26 February 2012 at 1:32 PM
T=2 in my case (0'14"). The T=1 case takes about 50"longer, and therefor reads 1'04"
@WandW: dont press cancel, just finish the initial render. Your T=1 was 1'37" while T=2 lasted 0'38". Values make perfect sense.
The observation on raytraced shadows is interesting, as compared to shadow maps: better=faster in this case.
- - - - -
Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.
visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though
aRtBee posted Sun, 26 February 2012 at 1:57 PM
so better = slower still. That's a relief.
- - - - -
Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.
visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though
bevans84 posted Sun, 26 February 2012 at 4:26 PM
i7 2600, 3400 mhz, 8 gig ram 19s PP2012
It started at 3800 mhz (turbo boost) settled down to 3500mhz for most of the render, and sped up to 3700 towards the end
JimTS posted Sun, 26 February 2012 at 4:50 PM
Unzipped obz"s
Config i7 970 w 24G ram
Poser 7 T2 render 1:43
Pro 2012 T2 render 0:24
A word is not the same with one writer as with another. One tears it from his guts. The other pulls it out of his overcoat pocket
Charles Péguy
Heat and animosity, contest and conflict, may sharpen the wits, although they rarely do;they never strengthen the understanding, clear the perspicacity, guide the judgment, or improve the heart
Walter Savage Landor
So is that TTFN or TANSTAAFL?
Eric Walters posted Sun, 26 February 2012 at 11:17 PM
Your image is very amusing. No P7 so I cant join the race!
Quote - Like pjz99, I built my machine 10 months ago and the only quad available was the QX6700. But
if I were going to built today, I'd go for the QX9650. Only because it's a 45nm architecture chip,
and the extreme chips are unlocked (better for overclocking). Now if money was an issue, I'd
wait for the regular quads to come out with the 45nm, these 65nm chips run hot, way HOT. But
all in all, I'm very happy with what I have now. Haven't been able to bring it to it's knees yet. And I
love loading up scenes. 8-)
prixat posted Mon, 27 February 2012 at 2:17 AM
Where can I find a timer in Poser Debut?
regards
prixat
LostAlien posted Mon, 27 February 2012 at 2:33 AM
Michael,
Thank you for your e-mail.
I believe my results are sound and are valid for the following reasons:
I use all the poser libraries from several previous versions of poser as well as 2012
I had no error or warning messages
All characters in the scene were skinned and the render was complete
The scene looked good to me
I have been using poser for over 12 years and am very familiar with the product
I have just opened Poser, added both male and female G2 tennis shoes
This worked without issue and rendered perfectly... See attached image.
If you are having issues with missing components and have legitimate copies of previous versions of Poser, I would suggest using your backups to restore your runtimes and include them in the list of libraries.
If you don't have backups available, install the earlier versions on another computer, create a backup of the Runtime then deinstall poser on that computer.
You will need to do this for each version.
Do not leave Poser installed on another computer unless you are licensed to do so.
If you do not have another computer available, you could try the following (at your own risk):
Take a backup
Full image if possible or copy the poser folder to another location.
Remove Poser 2012
This will be restored as it is in step 8 below.
Install earliest version of poser that you have
This will create the runtime folders with the missing content.
Make a copy of the runtime in another location
This is done so that you can use them once Poser 2012 is restored.
Remove this old version of Poser
Do this by using the uninstall, do not just delete the folder.
Repeat steps 3 to 5 for all your versions
This will result in all runtimes being restored and saved.
Reinstall Poser 2012
This will be a "Vanilla" build of Poser 2012
Overwrite the new Poser 2012 folder with the one taken in step 1 above
This will restore your poser folder to how it was when you removed it...
You should now be back to where you started but with all your runtimes
Add them to your library list in poser.
Note that I have not performed these steps as I use my backups. If you undertake these actions, you do so at your own risk. This is, in my opinion, a valid way to do this although you might want to check with SmithMicro to see if it would cause any issues or if they have a better solution.
You mentioned that the following files were missing:
On my computer, these are currently in my Poser 7 libraries so, unless I moved them from somewhere else, I assume that these came with Poser 7? Not sure, don't really care - They are there.
I have 5 libraries in my version of Poser 2012 - 3 are from earlier versions of poser (Poser 7, Poser 2010 and Poser 2012) and the others I created myself to keep my own content separate.
Having multiple libraries is a great feature of Poser.
Thank you for your concern.
I hope that this helps you.
LA
prixat posted Mon, 27 February 2012 at 4:52 AM
I got a message window asking me for those specific files.
I noticed Debut has them installed as OBZ and the P7 scene was looking for OBJ.
regards
prixat
WandW posted Mon, 27 February 2012 at 7:07 AM
If you point the request for the shoes' obj to the clothing for Simon and the pants to Sydney clothing (both under libraries/props) and select the appropriate obz, it will load fine.
One additional note; I have a scene with a lot of transparency and an SSS skin texture. In PP2012 with my Phenom X6 at 3200 MHz it rendered in 57' 35". Bumping it up to 3500 MHz cut the render time to 54' 2", which is about a 5% decrease in render time for a 10% increase in processor speed.
Max CPU temperature was 52 C at the higher speed, as compared to 46 C at stock speed...
----------------------------------------------------------------------------------------
The Wisdom of bagginsbill:
"Oh - the manual says that? I have never read the manual - this must be why."
aRtBee posted Mon, 27 February 2012 at 7:51 AM
@WandW
did the CPU ran at full 100% in both cases? Memory should have matching speeds too.
you can check the real net clock speeds with TMonitor from cpuid.com (freebie).
Thanks to Turbo Boost Technology, clock speeds are dynamic and temp dependant. So if your 3200-setup actually ran at 3400, and your 3500 setup could make 3600, then that's just a 5% diff. Or: overclocking requires over-cooling to get the real benefits out. Although your 52C is quite low, actually. PhenomX6 is quite nice.
- - - - -
Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.
visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though
WandW posted Mon, 27 February 2012 at 10:08 AM
When Poser renders, it uses all 6 cores at 100% of the base CPU clock speed. Turbo Boost doesn't kick in because all of the cores are being used.
At 3500 I'm running at 1.32 Volts at full load, compared to 1.26 V at 3200.
EDIT I should add that I could overclock the bus as well, but since it is an unlocked chip It's simpler to tweak the multiplier...
----------------------------------------------------------------------------------------
The Wisdom of bagginsbill:
"Oh - the manual says that? I have never read the manual - this must be why."
aRtBee posted Mon, 27 February 2012 at 10:59 AM
Individual threads however perform different, turbo off make most (!) of them run at 3464.5 (next image), but with turbo on none (!) of them made the full 4GHz (image after that). I am loosing some 4% there.
So, run TMonitor with the -T option (threads), and find out if your 3500-setting really delivers.
I'll stop this now, as we are sort of sidetracking this forum thread.
- - - - -
Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.
visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though
aRtBee posted Mon, 27 February 2012 at 11:00 AM
- - - - -
Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.
visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though
aRtBee posted Mon, 27 February 2012 at 11:01 AM
- - - - -
Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.
visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though
WandW posted Mon, 27 February 2012 at 3:29 PM
I have Turbo mode disabled at 3500. I looked at the CPU utilization with AMD Overdrive, and its charts look much like yours. The difference could be memory timing related, as 3200 is a multiple of the 800 MHz memory speed...
----------------------------------------------------------------------------------------
The Wisdom of bagginsbill:
"Oh - the manual says that? I have never read the manual - this must be why."
usamike posted Sun, 20 September 2015 at 2:36 AM
updated on 20 september 2015 since i update my configuration hardware, i redo this old test (not perfect, but the only poser benchmark on the internet!)
Please post your result with this form to help me tu update the tab array
V | T | Render TIME | thread | CPU | MHz | RAM | OS | User & small comment
4 | 2 | 0"11s | 12 | Ci7-980X | 3300 | 12GB | W7 64 bits | usamike (poser p2014 and SSD evo850)
it is awsome to see that poser pro 14 make my render quicker than the higher cpu (i7 990x) with older poser pro release (pp2012)
For whom wants to redo the test, i remember links (with textures fixed):
for poser pro 2012 and 2014 only users : http://www.michael-remy.fr/forum/pro2012.zip
for poser pro 2010 users : http://www.michael-remy.fr/forum/pro2010.zip
for other poser (7,8 maybe 6) users : http://www.michael-remy.fr/forum/poser7.zip
Legend :
v (poser version) :
4 : poser pro 2014
2 : poser pro 2012
0 : poser pro 2010
8 : poser 8
7 : poser 7
p : poser pro (release before the 2010 one)
And T means 1 or 2. When first render (load texture to disk) then it is 1, and when second render next, it is 2 (texture are already preload in memory)
Please post your result with this form to help me tu update the tab array
V | T | Render TIME | thread | CPU | MHz | RAM | OS | User & small comment
4 | 2 | 0"11s | 12 | Ci7-980X | 3300 | 12GB | W7 64 bits | usamike (poser p2014 and SSD evo850)
V | T | Render TIME | thread | CPU | MHz | RAM | OS | User
4 | 2 | 0"11s | 12 | Ci7-980X | 3300 | 12GB | W7 64 bits | usamike (poser p2014 and SSD evo850)
0 | ? | 0"14s | 12 | Ci7-990X | 4000 | 24GB | W7 64 bits | Artbee (poser p2010 and SSD)
2 | ? | 0"20s | 8 | Ci7-2820QM | 2300 | 8GB | W7 64 bits | Lost Alien (maybe missing texture)
2 | 2 | 0"22s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
0 | 2 | 0"25s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 2 | 0"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 2 | 0"37s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
2 | 2 | 0"38s | 6 |Phenom X6 | 3200 | 16GB | XP 64 bits | WandW
7 | 2 | 0"47s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 1 | 1"11s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 1"13s | 4 | Q6600 | 2400 | 8GB | Vista64 bits | svdl
0 | 1 | 1"15s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 1 | 1"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 1 | 1"37s | 6 |Phenom X6 | 3200 | 16GB | XP 64 bits | WandW
2 | 1 | 1"55s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
7 | ? | 1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAn
7 | 1 | 2"01s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 2 | 02:00 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
8 | 2 | 2"06s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
7 | ? | 2"11s | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
7 | ? | 2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
7 | ? | 2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
7 | ? | 2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits |ChippsyAnn
7 | ? | 2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPa
7 | ? | 2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
8 | 1 | 2"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
7 | ? | 3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunky
7 | ? | 3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfre
7 | ? | 3"40s | 4 | Q6600 | 2400 | 4GB | XP 32 bits | gsfcreato
8 | 1 | 03:45 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfre
7 | ? | 4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
P | 2 | 5"20s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
P | 1 | 5"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
7 | ? | 5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
7 | ? | 6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
7 | ? | 6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfr
7 | ? | 7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
7 | ? | 8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
usamike posted Thu, 28 December 2017 at 3:06 PM
updated on 27 december 2017 since i update my configuration software, i will redo this old test (not perfect, but the only poser benchmark on the internet!)
Please post your results with this form to help me to update the tab array
V | T | Render TIME | thread | CPU | MHz | RAM | OS | User & small comment 4 | 2 | 0"11s | 12 | Ci7-980X | 3300 | 12GB | W7 64 bits | usamike (poser p2014 and SSD evo850)
it is awsome to see that poser pro 14 make my render quicker than the higher cpu (i7 990x) with older poser pro release (pp2012)
For whom wants to redo the test, i remember links (with textures fixed):
for poser pro 2012 and 2014 only users : http://www.michaelremy.fr/pro2012.zip for poser pro 2010 users : http://www.michaelremy.fr/pro2010.zip for other poser (7,8 maybe 6) users : http://www.michaelremy.fr/poser7.zip
Legend :
v (poser version) :
11: Poser Pro 2011 4 : poser pro 2014 2 : poser pro 2012 0 : poser pro 2010 8 : poser 8 7 : poser 7 p : poser pro (release before the 2010 one)
And T means 1 or 2. When first render (load textures from hard drive) then it is 1, and when second render next, it is 2 (textures are already preload from memory or cached)
Please post your result with this form to help me tu update the tab array
V | T | Render TIME | thread | CPU | MHz | RAM | OS | User & small comment
V | T | Render TIME | thread | CPU | MHz | RAM | OS | User
4 | 2 | 0"11s | 12 | Ci7-980X | 3300 | 12GB | W7 64 bits | usamike (poser p2014 and SSD evo850)
0 | ? | 0"14s | 12 | Ci7-990X | 4000 | 24GB | W7 64 bits | Artbee (poser p2010 and SSD)
2 | ? | 0"20s | 8 | Ci7-2820QM | 2300 | 8GB | W7 64 bits | Lost Alien (maybe missing texture)
2 | 2 | 0"22s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
0 | 2 | 0"25s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 2 | 0"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 2 | 0"37s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
2 | 2 | 0"38s | 6 |Phenom X6 | 3200 | 16GB | XP 64 bits | WandW
7 | 2 | 0"47s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 1 | 1"11s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
7 | ? | 1"13s | 4 | Q6600 | 2400 | 8GB | Vista64 bits | svdl
0 | 1 | 1"15s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 1 | 1"33s | 8 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
2 | 1 | 1"37s | 6 |Phenom X6 | 3200 | 16GB | XP 64 bits | WandW
2 | 1 | 1"55s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 1"57s | 4 | QX6700 (over) | 3000 | 4GB | XP 64 bits | Magoo
7 | ? | 1"58s | 4 | Q6600 | 2400 | 8GB | XP 64 bits | ChippsyAn
7 | 1 | 2"01s | 4 | CI7-975 | 3300 | 6GB | W7 64 bits | usamike
8 | 2 | 02:00 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 2"07s | 4 | Q6600 (over) | 3200 | 4GB | XP 64 bits | ghonma
8 | 2 | 2"06s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 2"10s | 4 | QX6700 (over) | 3000 | 8GB | XP 64 bits | pjz99
7 | ? | 2"11s | 4 | E6850 (over) | 3300 | 2GB | XP 32 | Dalroi27
7 | ? | 2" 11s | 4 | QX6700 | 2670 | 4GB | XP 64 bits | Magoo
7 | ? | 2"15s | 4 | T7700 | 2400 | ???? | Vista 64 | adom
7 | ? | 2"23s | 2 | Q6600 | 2400 | 8GB | XP 64 bits |ChippsyAnn
7 | ? | 2"25s | 4 | Core2Extreme?? | 2800 | 2GB | MacOS | MungoPa
7 | ? | 2"46s | 3 | Q6600 | 2400 | 4GB | XP 32 | xpac5896
8 | 1 | 2"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 3" | 4 | Q6600 | 2400 | 4GB | XP 64 bits | ghonma
7 | ? | 3"06s | 4 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"07s | 2 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 3"08s | 4 | E6600 | 2700 | 2GB | XP 32 bits | thefunky
7 | ? | 3"21s | 4 | 2x Opteron 285 | 2600 | 8GB | XP 64 bits | renderfre
7 | ? | 3"40s | 4 | Q6600 | 2400 | 4GB | XP 32 bits | gsfcreato
8 | 1 | 03:45 | 2 | Athlon 64x2 | 3000 | 8Gb | w7 64 | Ricky1
7 | ? | 3"56s | 2 | E6400 (over) | 2320 | 2GB | XP 32 bits | renderfre
7 | ? | 4"41s | 1 | AMD X2 6400+ | 3200 | 4GB | XP Pro 64 | UrbanArm
7 | ? | 4"43s | 2+HT | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
P | 2 | 5"20s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
P | 1 | 5"50s | 4 | Q9550 | 2800 | 4GB | XP 64 bits | Michael314
7 | ? | 5"40s | 4 | 2x AMD X2 5200+ | 2610 | 4GB | Vista 64 | tastiger
7 | ? | 5"50s | 2 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
7 | ? | 6"03s | 1 | T7700 | 2400 | 4GB | XP Pro 32-bit | UrbanAr
7 | ? | 6"31s | 1+HT | Pentium 4C | 3400 | 2GB | XP 32 bits | renderfr
7 | ? | 7"02s | 3+HT | P4 | 3200 | 2GB | XP 32 | xpac5896
7 | ? | 8"26s | 1 | Pentium D960 | 3600 | 2GB | XP 32 bits | usamike
willyb53 posted Thu, 28 December 2017 at 8:17 PM
I will play :D
V | T | Render TIME | thread | CPU | MHz | RAM | OS | User
2 | 2 | 0"14.5s | 24 | X5670 (2) | 2930 | 48 gig | win7 pro | willyb53
People that know everything by definition can not learn anything
bantha posted Sat, 30 December 2017 at 9:00 AM
V | T | Render TIME | thread | CPU | MHz | RAM | OS | User
11| 2 | 0"12.8s | 16 |Ryzen 7 1700 | 3000 | 16 gig | win10 pro | Bantha
11| 2 | 0"10.6s | 16 |Ryzen 7 1700 | 3800 | 16 gig | win10 pro | Bantha
I don't think that this scene is complex enough to get good results from a lot of cores.
A ship in port is safe;
but that is not what ships are built for.
Sail out to sea and do new things.
-"Amazing
Grace" Hopper
Avatar image of me done by Chidori.