I'm considering if I should get a 3GHz P4 or an Athlon 3200+ but I don't really know which is faster. So we created a program called Speedtester. If you own a 3GHz P4 or Athlon 3000+ or better, you can download "SpeedTester" from our homepage. Our faster PC is an Athlon 2100+ which completed the 50 million calculation test in 1.50 seconds. I'll be happy if you email me your results and specify which CPU you used if you have any of the above systems. Thanks.

I've just picked up a P4 3.06 laptop yesterday. My Desktop is an XP 2800+. Can I compare those 2?

What is the address of your homepage?

he he 0.53125 Seconds for 5 Million 1.050781 Seconds for 10 Million 5.269531 Seconds for 50 Million on my trusty AMD-K7 500MHz computer

0.1796875 secs---5 million 0.3710938 secs---10 million 1.884766 secs---50 million my machine is nowhere as fast as it was when it's empty...as it stands now i'm filled up about 98% and crazy cluttered..still backing up files so i can format. 1.8ghz pentium---2gigs ram---100gig drive

0.125 Seconds for 5 million 0.25 Seconds for 10 million 1.265625 Seconds for 50 million. I have a P4 2.4ghz.

Pentium 4 3.06 GHz Notebook: 0.9882813 Seconds - 50 Million 0.1914063 Seconds - 10 Million 9.765625E-02 Seconds - 5 Million

JVRenderer: 9.765625E-02 Seconds for 5 million..? Did your machine throw up an Error..?

Whats the URL for the site?

1Freon1, click on the name of d4500's name and go to the artist's page. It's listed at the bottom of the page. Ziggie, that is less than 1/10 of a second (E-02 is 10^-2) Any suggestion which download from MS provides the VBVM.dll executable that's required? I am curious how my 2.4 GHZ performs considering I'm running a windows emulator. eric

Figured it out Eric, but thanks. Athlon T-Bird 1.2Ghz 5 - 0.21875 10 - 0.4492188 50 - 2.28125

Pentium 4 1.7 50--1.699219 10--0.3513625 5 --0.1679688 That's cool. I'm thinking of getting a 3Gigger myself. Love to know how much faster it would be. I'll try the 2.4 later-- when I can move this, need a new wireless card. Emily

If you get a Pentium, make sure you get one with the 800mhz Front Side Bus and a motherboard with the Intel 865PE Chipset. I'm thinking of upgrading to a Pentium 2.6/800. This article has a good performance comparison between current Intel and AMD Processors. http://www17.tomshardware.com/cpu/20030521/index.html This site is a good source of up-to-date price comparisons of computer hardware. http://www.pricewatch.com This site has evaluations of various mail-order vendors. http://resellerratings.com Visit them all and draw your own "bang-for-the-buck" conclusions.

Attached Link: http://www.sisoftware.net/index.html?dir=dload&location=sware_dl_x86&langx=en&a=

Best result 2.8 sec. @ 50 Million Worst result 3.4 sec. @ 50 Million Athlon 1GHz Slot A. Of course, to do a thorough benchmark, you really need something like Sisoft Sandra.

THanks for helping out. The speedtester isn't a benchmark utility... but it does test how fast the CPU can process a single repetitive task. 3GHz: 0.988 seconds @ 50Million, makes it 53% faster than my 2100+ (1.52). I am eager for someone with an Athlon 3000+ or better to get their results. Normally, it takes a minute or so to do a rendered scene. However, I had six HiRes models in a scene. It took over 7 minutes to render... so hence the desire to get a new CPU.

Finally found the file. P4 3.06 0.984375 @ 50 Million 1.4375 @ 50 million X 2 copies at once I'm going to hang on to the little program for when I get new machines. :-) Thanks!

I have a XP 2800+ Desktop at home, but I am out on a business trip and won't be home till July. By that time this thread is probably buried.

What's the calculation you're doing in the loop? I'd be interested in comparing how well the VB compiler optimizations work and also VB vs VC++.

P4 3.06 GHz 1gig Ram 50 Million 0.9394531 Seconds 10 Million 0.1875 Seconds 5 Million 0.09375 Seconds asus P4G8X Deluxe MB

Thanks for your input. I like this software because it's a good comparsion for my PCs. Here's the code: Command1_click 'Set the variable x depending on user's option setting. ' x = 50000000 'Default 50 million if option2.value = true then x = 10000000 if option3.value = true then x = 5000000 '------------------Timer has started--------------------- a = timer ' get the current reading from the timer for y = 1 to x ' repeat from 1 to x next y b = timer ' get a new timer reading '------------------Timer has stopped--------------------- c = b - a ' get the time elapsed label1.caption = c 'display results I expect that it runs faster in C++. It runs faster in VB (1.20 @ 50M) than as an EXE (1.52 seconds @ 50M) - Athlon 2100+. Have fun. Does anyone know a friend with an Athlon 3000+ or better? I'd be happy to see the results compared to the P4 3GHz. To calculate the speed percent between 2 PCs.... Slower Timer / Faster Time = % faster 1.52 (2100+) / 0.98 (3GHz) = 1.55 or 55% faster. Therefore, the 3GHz is 55% faster than the Athlon 2100.

Ummm...Hasn't there been a good bit of fury over the fact that Sisoft has reportedly optimized their software in favor of the Intel products? And not to rain on anyone's parade, but that -only- shows you a single operation execution, and that does =not= equate to anything like real world performance. This doesn't excercise either the L1 or L2 cache, and how they perform can have a tremendous effect on the results. It doesn't seem complex enough to stress the branch predictor and prefetch predictor systems either; another area where there clock cycles are secondary to architecture. Then you have to take into account frontside bus latency, access time for both the ram and the hard drive(s), what OS is controlling all of this and what it is doing with and to the system resources, PCI and AGP bus latencies, interrupt and DMA requests, possibly even thermal conditions in the case and room. All this really shows is that a 3 ghz P4 is clocking %55 faster than an XP-2100 (which a glance at a spec sheet will tell you anyway).

Since this particular benchmark fits entirely in 256KB L2 cache, scaling should be linear with cycle time. Even the 3200+ Barton should be slower than the 2800+ Thoroughbred since the Barton runs at only 2.2 GHz versus 2.25 GHz for the Thoroughbred. Dale LaRoy Splitstone

Attached Link: http://www.mvps.org/ccrp

I kept getting wildly different results from basically the same code until I realized that you must not be typing your variables. By default in VB, you get variants which take up 16 bytes each as opposed to 4 bytes for a long integer or a single. There is also overhead for doing type conversion from a variant to the desired type. The difference is quite dramatic. I get sub 200 Millisecond times (averaging 10 iterations). The code below uses a hi resolution timer object you can get at the link. I also tested using the timer function as you did, which yields slightly faster results but probably not as accurate. This doesn't invalidate your relative performance numbers but it shows why you always want to use typed variables. 155 Milliseconds vs 2.8 Seconds. I did turn on all the compiler optimizations for this executable. Not using any optimizations is only slightly slower - a few milliseconds. '--------------------------------------------------------- Dim objStopWatch As ccrpStopWatch Dim lngCount As Long Dim intIterations As Integer Dim lngTotalTime As Long Dim lngAvgMs As Long Dim strResults As String Me.Caption = "HiRez Timer 10 Iterations" Set objStopWatch = New ccrpStopWatch For intIterations = 1 To 10 objStopWatch.Reset For lngCount = 1 To 50000000 Next lngCount lngTotalTime = lngTotalTime + objStopWatch.Elapsed Next intIterations lngAvgMs = lngTotalTime / 10 strResults = "lngCount: " & lngCount - 1 & vbCrLf strResults = strResults & "Total Time: " & lngTotalTime & " Milliseconds" strResults = strResults & vbCrLf & "Avg. Time: " & lngAvgMs strResults = strResults & " Milliseconds (" & (lngAvgMs / 1000) & " Seconds)" lblResults.Caption = strResults Set objStopWatch = Nothing '---------------------------------------------------------

For some reason, when I try to post bota a link and a picture, it never works :-)

Attached Link: http://www.pcworld.com/news/article/0,aid,110484,pg,2,00.asp

Amd posted benchmark results for the opteron where a 1GHz itainium outruns a 2GHz xeon (amd.com). The speedtester was just a simple program for my own use to compare my PC with others (as in my PC can do this... how fast can yours do it?). PCworld (link) benchmarks the 2.17GHz athlon (3200+)outruns the Pentium 3.06GHz. Does that mean the athlon is faster? Not necessarily. One results may say the athlon is better, the other might say the P4 is better.. but I think I'll stick with which one counts.

That's the whole thing about benchmarking now. Intel and AMD have taken different approaches to how CPU's handle work. A P4 using a program that is optimized for SSE-2 and with HT active will certainly kick an Athlon's ass. But 96% of the software out there doesn't know what either of those thing are. Most code is still plain old X86, with SSE, 3D-NOW! and other extensions bolted on. Without those, you get into the meat of the design philosophy. Clockspeed isn't the easy win issue it once was, anymore than engine RPM means jack. If you don't have a properly designed torque converter and transaxle, you might have a 20 cylinder gold plated monster that can sustain 12,000 RPM's but a stock Yugo could still outdrag you. AMD has been keeping pace, performance-wise, with Intel since it reclaimed the clockspeed crown, with design philosophy; namely the length and width of the pipe, the design of the prefetch, branch predictor, and cache controlling logic. And the XP-3200+ isn't a 'faster' chip. It's a better chip, that does more on less cycles per second. As for the Itanium vs. the Xeon, I believe the former has a much larger L1 and L2 cache. The less you have to go to the external buses, the 'faster' an operation becomes.

As d4500 says, this is just an interesting little comparison. People love to talk about how fast a given car will go from 0 to 60, despite the fact that unless you're a drag racer, it's not very relevant to real life situations. If it wasn't so much of a hassle, I'd drag out my old Z80 based Timex-Sinclair 2068 and see what it would do on this test - just for grins. I think all the tech details of microprocessor architecture are interesting but the only real way to determine which CPU will render a scene how fast is to try it. All the specs and synthetic benchmarks in the world are no substitute for running the real application in your typical production environment. Techie tidbit: Each new Opteron server from Newisys includes an IBM PowerPC chip service processor running Linux that monitors the system. Buy now before SCO makes it illegal!

Boy, I could really use the Athlon 3000+/3200+ results about now... anyone???