PiFast 43 question

[Liquid3D]

I downloaded Pifast 43, and entered the following commands;
0
0
10000000
1024
0

I checked the Internet, but nowhere could I find specific instructions on what values to enter. The "log" file from Hexus happened to give me the values, but I ran the Pifast 43 thinking it was the "new" up-to-date version? Now I notice three differences. First, everyone seems to be using version 41. Second not too many seem to enter any value after "Enter FFT size in k :", and finally where it displays Total Physical Memory Allocated = 61373 " In the version I ran it's 61757 kilobytes? See below. Also the screen dissapears before I can take a snapshot. I was lucky below by taking the Pic as the program tallied results, but i would never have time to asttach a WCPUID screen as well?

[hipro5]

Pi-fast V.4.1 is picked for the benchie and I think it won't change with the new version cos AMD's don't want to.......It's a lot faster with a P4 rig and that's the reason they don't want to change the benchie cos the difference between an AMD and a P4 would be a LOT BIGGER.......

I can fix a "standard" bench for you if you want to......

[Liquid3D]

OK here'a another question if ya'll don't mind. Why are we entering 1024 kilobytes under the amount of of "Physical memory"? Does this require a gigabyte in the system? I happened to have 512MB in, I'll throwanother matching pair in , but would entering 512k be the same? And how do you take the snapshot without PiFast dissapearing? As soon as it's finished it just ends? And the MSDOS window closes so quickly I don't have time to take a snap shot, leta alone take one with WCPUID next to it as many do? I even left Paraben (screen shot taker) running minmized ready to photo, with WCPUID minimized as well, to open quickly, but wouldn't this slow the benchmark?

These may be silly questions, but I'd really appreciate the help? Here's Pic below, where I ran under 512k because that's what's in the system? I'm calculating wrong aren't I? (Doh) 512 kilbyte, and 512MB are two different amounts, and wheter use you have 512MB or 1-GB of RAM, the benchmark is measuring the CPU's Math co-processor not memory? Is that right?

[Liquid3D]

Now here's my latest result and again i had to leave WCPUID, and PARAben screen capture running in order to take the snapshot while PiFast was still open? At least "Total Allocation matches everyone's at 61372. But I notice almost everyone allows the ENTIRE cycle to complete and then take the snapshot? PLEASE HELP I'm beginning to loose it. (below is 250FSB 1:1 GolD3700 512MB, Abit IC7-G Beta BIOS revision 17);

[Gleep]

Download this , it contains the batch file everyone is running, which includes a pause command so you can screenshot it.

[Liquid3D]

Thank you Gleep, and welcome to the Xtreme. Hope you stick around because people who scan thread, and offer help are what makes these forums work. And that's no blue-smoke.

I found the batch, and ran it, and what depressing results. Of course I had DSL running off USB. Still then time is horrific isn't it? For 2.4GHz and 200FSB?

[hipro5]

I think that the link is with the V.1 witch is slower than the V.3 .........

[PMM]

All depends if your intending to send in a benchmark
becasue I'll reject any 4.3 results.

I use 4.1 because thats the way its always been to change
it will completely mess up the table and be unfare to all the
others.

[Liquid3D]

I want to reamin consistant with your comparison chart. I'm certainly not coming along asking you to make any changes. But tell (because you replied after I'd thought I had the right version) me is this the version? It's the batch file Gleep linked me to, and I didn't have to enter any values to run it. It paused upon completion so I could take a screenshot. So this is the version right?

[Gleep]

The version I linked to was pulled directly off of PMM's page, I copied the url from his "download" button. So you are using the offical version, v4.1.

[kromosto]

these are my scores at p42.6@3.3316

[Gleep]

Please convert that screenshot to a jpg, this bmp stuff takes forever to load. ack, no wonder it's 3MB.

[kromosto]

ok done!

[LikwidKool]

Thanks for making this thread liquid, as I had no idea if I was running the right version with the correct switches. I am still confused though as my score seems like it might be a little high.

BTW does everything look correct in my screenie?

[Liquid3D]

Thanks LiKwidKool. From what I see your using the right version. The correct version requires no input, while the Ver.43 requires all the values be entered. And I'd say that's a very good score, although being new at this I'm not positive.

I'm wondering if this is solely based upon the Processor, or if RAM has any influence on the results. Your latency's being what they are seem's to have improved you timw significantly. I'm less then a 100MHz slower, yet you have 3-sec on my score. I think even 3-sec is significant in this test.

[Gleep]

Memory speed and timing effect pifast, the difference between tight timing and relaxed timings on my P4C800 can be 1-2 seconds at the same cpu/memory speed.

[LikwidKool]

Thanks gleep, and liquid!

[Liquid3D]

hey there LiKwidKool I notice yiou have what I call the "dreamaker/heartbreaker" system with the 2.4C Malay, Abit IC7, and Kingston Hyper-X3500 (currently the undisputed king of 875 bandwidth) are you the person I saw the 7000MB/s bandwidth benchmark from? If so could you send me a pic, I having a watercooled micro-PC designed/built and need to submit some proof as tio the Hyper-X3500's track record.

By the way I ran the benchmark again this time using Corsair Twin-X4000 is this a decent score in the scheme of things?

[LikwidKool]

sorry buddy, I wish I was the one with the 7000k. Best I have done is 6300. I get shut down at 287 5:4 so far. A little dissapointed considering I did the vdimm mod too. I just ordered some Corsair pc3500 v1.1(one with winbond chips) and gonna see if this does any better than the hyperx. Mine doesn't like going above 233 at 2-6-3-2 for some reason. Then again if I could feed more than 3.0v into it I might do better. Thing is I need a stable VTT cause above 3.0v vdimm it gets hairy. I need to start playing around with bios' again, cause I swear my system was the most stable, and clocked highest with the first bios that gave GAT. I think it was 15? I keep decent screens of all my best benches and lable them according to bios and all my good ones seem to come with 15 or beta 15. After the Back To School rush dies down(I manage a 4mill/year Sporting Goods Store), I am gonna do some testing.

BTW 5905 is pretty good for 270 wouldn't ya think?! If I do things right I can sometimes break 6k at 275. As far as the pifast I am not good with telling results, but it looks in line with what I got at 283.