wPrime 2.00 beta

[wwwww]

In an effort to solve the problem of unbeatable scores achieved with older wPrime versions, I've completely re-done the calculation and threading section of wPrime in an effort to bring the score on par with or slightly faster than v1.53.

I could use some testers to compare scores with older versions, in particular on higher processor count SMP systems as I've completely redesigned the threading model so I'm not sure of the effect it'll have on 8 core + systems (only have quads here to test on).

- Wprime 2.00 Beta6: http://www.wprime.net/?q=download&f=wprime_200_beta6.zip
- Wprime 2.00 Beta3: http://www.wprime.net/?q=download&f=wprime_200_beta3.zip
- Wprime 2.00 Beta2: http://www.wprime.net/?q=download&f=wprime_200_beta2.zip
- Wprime 2.00 Beta1: http://www.wprime.net/?q=download&f=wprime_200_beta1.zip
- Wprime 2.00 Beta0: http://www.wprime.net/?q=download&f=wprime_200_beta.zip

Compare with:
- Wprime 1.55 (current stable): http://www.wprime.net/?q=download&f=wprime_155.zip
- Wprime 1.53: http://www.hwbot.org/blog/wp-content/wprime_153.zip
- Wprime 1.43: http://www.hwbot.org/blog/wp-content/wprime_gui143.zip

Basically would like to know, too fast or still too slow?

Note that you won't be able to submit scores at this stage.

Thanks :thumpsup:

[killer55555]

I think it's about 2.5 sec faster(the v2).I have run it twice to make sure of it.

V2 gave me 30.6 sec and the V1.55 33.1 sec

[Advanced]

About 2,9 sec. faster than v.1.55 on my X2 5200+.

http://www.abload.de/thumb/unbenannt2p4.png (http://www.abload.de/image.php?img=unbenannt2p4.png)

[WoOx3r[Pt]]

Were i found the biggest differences on timings where on amd k6 machines so i've tested version 1.53 1.55 and 2.00 on a stock amd k6-2+ 570mhz to see the differences.
-
version 1.53 - 441.184
version 1.55 - 528.781
version 2.00 - 353.238
-
image - http://img187.imageshack.us/my.php?image=wprimeversionsperformanxn7.jpg

The version 2.00 didn't detected the cpu as it should but version 1.53 and 1.55 did that perfectly
-
image - http://img87.imageshack.us/my.php?image=wprimeversionshardwaredqk3.jpg

As we all can see the overall look of wprime 2.00 it's much more attractive than it's older fellas
-
image - http://img529.imageshack.us/my.php?image=wprimeversionsappearancpf6.jpg

As many people know the old hardware like k6 processor based machines isn't known by making very accurate scores BUT this version of wprime is a little bit more accurate than the others but all of them showed a maximum variation on timings of 0.075% wich is very impressive considering that SuperPi sometimes give me a 0.25% variation under the same conditions.

Overall evaluation - If this wPrime version identified the CPU as it should it would deserve a 110% approval by me but it doesn't, so i give it only 100% approval xD because wPrime was made to test the perfomance, not to detect CPU's, for that we have CPU-Z xD but if this small error gets corrected i think that this isn't just one more version of wPrime, i think this is THE VERSION of wPrime

Thanks for making this new version and for the huge efforts to do it.

Please let this version or an improved/corrected one be the official, you guys will be making the delights of many overclockers around here ;).

[j1g1t]

I downloaded this version. Tomorrow I'll post my opinion about this beta

[wwwww]

Thanks for the feedback - The approved hwbot version is up to Frederik, but with those types of scores, I'd say I need to slow it down a little - I'm aiming for it to be on par with version 1.53.

[wwwww]

Ok how about this:

http://www.wprime.net/?q=download&f=wprime_200_beta1.zip
Should be slightly slower - how does it perform on the K6?

[MrCape]

Please hit us with a big red warning when this new version becomes a requirement. I put a good amount of LN into v1.55 testing a couple slow cpus. :)

[WoOx3r[Pt]]

k6-2+ 570mhz

version 1.53 - 441.184
version 1.55 - 528.781
version 2.00 - 353.238

version 2.00 beta 1 - 382.79

version 2.00 beta 1 is more closer to the 1.53 than the version 2.00 beta but it's still booming way faster than version 1.53 xD

[jabski]

0.136 difference here.

http://img115.imageshack.us/img115/4517/wptestpl1.th.jpg (http://img115.imageshack.us/my.php?image=wptestpl1.jpg)

[jmke]

is it possible to set the amount of threads in Wprime 2.0 ? in the older one you could adjust the amount in the "advanced" menu

1.43 1.53 1.55 2.0b0 2.0b1 results below:
345

[Advanced]

Again comparison on X2 5200+

http://www.abload.de/thumb/unbenanntwzg.png (http://www.abload.de/image.php?img=unbenanntwzg.png)

Perhaps I will test the different versions on my K6-2 450mhz this weekend.

[toutatoc]

I've tested all versions on my laptop @Stock

AMD Mobile 3400+@1.8GHz 512Mo DDR2

- 1.43 102.485s http://img373.imageshack.us/my.php?image=wprime143yx1.png
- 1.53 104.844s http://img137.imageshack.us/my.php?image=wprime153zb2.png
- 1.55 97.906s http://img145.imageshack.us/my.php?image=wprime155mr7.png
- 2.00 beta0 93.047s http://img373.imageshack.us/my.php?image=wprime200beta0ct4.png
- 2.00 beta1 93.062s http://img145.imageshack.us/my.php?image=wprime200td4.png

No problem for detection ;)

Next version ,that's sub 90s for the same frequency :D

[KTE]

Not a CPU most of you would check out with wP but oh heck... its the thought that counts :p

Vista Ultimate 32b
1.43 - 1.53 -1.55 - 2.0 - 2.0b1
34.017s - 33.432s - 33.41s - 31.788s - 33.025s

http://img502.imageshack.us/img502/3547/143ze8.th.png (http://img502.imageshack.us/my.php?image=143ze8.png) http://img527.imageshack.us/img527/5099/153ku8.th.png (http://img527.imageshack.us/my.php?image=153ku8.png) http://img526.imageshack.us/img526/9969/155eg6.th.png (http://img526.imageshack.us/my.php?image=155eg6.png) http://img257.imageshack.us/img257/2029/20bxt8.th.png (http://img257.imageshack.us/my.php?image=20bxt8.png) http://img516.imageshack.us/img516/5306/20b1le9.th.png (http://img516.imageshack.us/my.php?image=20b1le9.png)

I can add XP SP3 32b if anyone wants (dual boot).
Also I've just got rid of some P4, K6 and K7 but still have one last P4 2.8G left I'll try adding.

[wwwww]

is it possible to set the amount of threads in Wprime 2.0 ? in the older one you could adjust the amount in the "advanced" menu

1.43 1.53 1.55 2.0b0 2.0b1 results below:
345

Yeah, just hit ~, type
ft x
where x is the thread count
<enter>
then
gui
<enter>

I might put a thread change option in the interface before release though.

[wwwww]

http://www.wprime.net/?q=download&f=wprime_200_beta2.zip

Still too fast, this should be slightly slower again.

[jmke]

I might put a thread change option in the interface before release though.

that would be cool; as hyperthreaded CPUs do see a nice benefit :)

Under XP E8400 @3ghz
- 1.55: 29.063
- Beta1: 28.594
- Beta2: 31.015 (one of the runs it didn't show value after .xxx although log showed 31.xxx)

Win2003 - Athlon 64 2.2Ghz (single core Winchester)
- 1.55: 85.797
- 2.0B1: 80.594
- 2.0B2: 84.953

[der8auer]

32m

2x AMD FX-74 @ 3000MHz

Beta : Doesn't detect 4 Cores - Just 2
Beta 1: 14,08s
Beta 2: 14,099s
v.1.55: 13,141s

[wwwww]

Hmm , seems too slow.

http://www.wprime.net/?q=download&f=wprime_200_beta3.zip

Another thing, 2.00 will be perfectly consistent between versions (as the calculation modules won't be recompiled).

[der8auer]

Yes beta3 is faster

32m

2x AMD FX-74 @ 3000MHz

Beta : Doesn't detect 4 Cores - Just 2
Beta 1: 14,08s
Beta 2: 14,099s
Beta 3: 13,341s
v.1.55: 13,141s

[jmke]

Under XP E8400 @3ghz
- 1.55: 29.063
- Beta1: 28.594
- Beta2: 31.015
- Beta3: 28.626

[loonym]

Odd that the zip package says beta3 but the file version inside says beta1
http://lakesidepc.com/wptest.jpg

[jmke]

Beta2 also says Beta 1;) he just forgot the rename

[George_o/c]

There is something weird going on here ... But I can't figure out exactly what can be blamed for that ...

I did my testing in two different Win XP installations (sorry gpr, but my Vista + SP1 nLited installation wouldn't run wPrime :D - have to create a new one ;)) ... One of them is my crappy one year installation (win xp pro + sp2), and the other one a fresh Win XP pro + sp2 nLited one ...

In the first one I used Olive Luna theme (from my tests, it's giving me a 0.5 - 0.7sec boost against the Classical one), and in the second one the Windows Classical Theme ...

Now, as far as the first one is concerned ... Take a look :

v1.43

http://img92.imageshack.us/img92/6682/v143dx4.th.jpg (http://img92.imageshack.us/my.php?image=v143dx4.jpg)

v1.53

http://img329.imageshack.us/img329/6449/v153wf4.th.jpg (http://img329.imageshack.us/my.php?image=v153wf4.jpg)

v1.55

http://img50.imageshack.us/img50/8420/v155tu9.th.jpg (http://img50.imageshack.us/my.php?image=v155tu9.jpg)

v2.00 beta 3

http://img329.imageshack.us/img329/8044/v200ur1.th.jpg (http://img329.imageshack.us/my.php?image=v200ur1.jpg)

after five runs, with v2.00 beta 3

http://img211.imageshack.us/img211/7479/weirdthingrk3.th.jpg (http://img211.imageshack.us/my.php?image=weirdthingrk3.jpg)

Then I changed to an other hdd, the one with XP + SP2 nLited (nLited setup) ... classical theme used ... Now I have the photos on my other hdd (so can't post 'em), but I have some videos ...

Same mobo, bios settings, rams, ram timings/sub-timings, cpu frequency ...

v1.55 = 85.187sec
v1.53 = 84.391sec
v1.43 = 81.360sec
v2.00 beta 3 = 80.422sec

What the heck is going on here ???

I have the videos and the photos, I can upload them tomorrow if you want ;) In the meanwhile, you can check out the boost I get from Olive Luna theme ... :p

P.S. Tweaks used ... Real Time priority ... only one ... ;)

//EDIT: Just read the thing about beta 2, beta 3 and how they are automatically re-named in beta 1 :p I've downloaded the beta 3 package, so it's beta 3 anyway :)

[wwwww]

Beta2 also says Beta 1;) he just forgot the rename

Nah I did it to confuse everyone

By the way, Realtime priority won't boost performance as the calculations are not done in the parent process (unless you do it to the children process).

[wwwww]

http://www.wprime.net/?q=download&f=wprime_200_beta6.zip

Can someone with a multi-CPU system test that this correctly sets the thread count (as in cores*processors - eg dual Clovers should run with 8 threads)

[jmke]

E6550 WinXP
- 1.55: 40.334
- 2.0 B6: 39.173

[cbjaust]

I ran each version 10 times on an underclocked Brisbane A62 X2 4800+
EDIT: Windows XP

AMD Athlon 64 X2 4800+ @ 2167MHz [HT Link @ 812.5MHz (3 x 270.8)]
G.Skill F2-8800CL5D-4GBPI. 4GB @ 270.8MHz (RAM divider set to DDR2 400)
Abit Fatal1ty AN9 32X (nForce 590 SLI - C51XE-MCP55XE)

Run wprime 1.43 wprime 1.53 wprime 1.55 wprime 2.00 beta6
1 37.375 39.172 38.577 38.828
2 37.468 38.921 38.593 38.859
3 37.468 38.750 38.609 38.781
4 37.390 39.000 38.609 38.859
5 37.546 38.641 38.702 38.844
6 37.421 38.828 38.562 38.891
7 37.281 39.000 38.625 38.781
8 37.531 39.000 38.609 39.000
9 37.312 38.813 38.593 38.906
10 37.672 38.625 38.609 38.921

mean 37.446 38.875 38.609 38.867
stdev 0.117023454 0.175161513 0.037490147 0.066392771

hope this helps! http://www.hwbot.org/forum/images/icons/icon14.gif

[jmke]

excellent tests :)

all those scores are very close to one another, I think issues with inconstancy is mostly with multi sockets & older machines

[knopflerbruce]

excellent tests :)

all those scores are very close to one another, I think issues with inconstancy is mostly with multi sockets & older machines

Nice work so far, but it's important that older systems can be benched - even if it means that there will be slightly more problems with newer rigs.

[wwwww]

See some of the K6 scores the first page, it's much faster on older setups.

[j1g1t]

Fantastic tool!
One of the best features is save results.

[beleno]

hi! tested on an old pc with celeron lga755@2.66ghz prescott G1

1.55=171.968 s.
2.0 beta 6=148.187 s.

if i can help with some test, i'm happy to do it. I can test with p4 s478, p4 mobile s479, sempron 754, athlon 939. Thank you

[jmke]

wow, that's quite a big difference

[knopflerbruce]

wow, that's quite a big difference

Correct me if I'm wrong, but we need to compare to 1.43 on older chips, not 1.55;)

[richba5tard]

Thanks for all the testing guys! Free hwboints for everyone! :D

[WoOx3r[Pt]]

version 1.43 - 438.393
version 1.53 - 441.184
version 1.55 - 528.781
version 2.00 - 353.238
version 2.00 beta 1 - 382.79
version 2.00 beta 2 - 417.48 * **
version 2.00 beta 3 - 386.275 *
version 2.00 beta 6 - 382.99 *

* beta versions with correct core detection
** beta version with the closest timings to version 1.43 BUT it gave me a 0.3% of maximum difference between some runs and that isn't so accurate as all the other 2.00 beta versions that gave me a difference of less than 0.05% between runs

[wwwww]

Thanks for that. All the 2.00 versions detect hardware information the same way, there must be other reasons why it doesn't detect correctly. Do you have CPUz running in the background? If so, you should close it before starting up wPrime.

Beta2 is too slow for modern CPUs, I'm thinking the 1/3/6 (they all use the same calc script) is most suitable as they still perform comparatively fast to 1.55.

[jmke]

Correct me if I'm wrong, but we need to compare to 1.43 on older chips, not 1.55;)

we also need to compare current 1.55 with this benchmark on newer CPUs, and the difference for that Celeron S775 is huge; 2.00b6 is currently too fast for old CPUs and Intel low end; while it's ok for newer AMD and multi socket

very hard to make it all match up. when can we expect beta7? :D

[WoOx3r[Pt]]

if a wprime version that it's faster than any other on every machines gets released it can have some huge differences sometimes but it's allways better than performing slower on some machines and faster on others because this way everyone is able to beat records and be beaten by the previous recordists and by the new overclockers.

i don't know how the calculation method is made on wprime but i had this idea that i don't know if it's possible but maybe it can help you out, to standardize the wprime timings maybe you can mix up the calculation methods of different wprime versions, i don't know if i'm saying something stupid or if you already done that but it's just a thought.

[jmke]

if CPU detected in CPUID list use 1.43? otherwise use 1.55 code? something like that?

[wwwww]

I don't think there's a problem with being too fast on older CPUs, as long as it's faster, as there aren't too many scores that'll be affected and it'll just require retesting.

What's important that results on modern day CPUs are still in line with version 1.55 and that 2.0 is faster than 1.43 which I think 2.0 succeeds at.

[WoOx3r[Pt]]

jmke that wasn't really my idea but it's a good idea too. i don't think it would be necessary to do that because as wwwww says as long it's faster it will only require re-testing, no more unbeatable scores because of version incompatibility issues. i hope one of this 2.00 versions of wprime gets approved ;).

[jmke]

I don't think there's a problem with being too fast on older CPUs,

version 1.43 - 438.393
version 2.00 beta 6 - 382.99

that's a whole lot faster in my book ;)

[der8auer]

Beta6 with 2x FX-74:

32m: 13,743s

[wwwww]

version 1.43 - 438.393
version 2.00 beta 6 - 382.99

that's a whole lot faster in my book ;)

Yes but older CPU's represent such a small portion of the scores, all it'll mean is a few benchers will need to re-bench. As long as the bulk of the scores (modern CPUs) are consistent, it should be fine.

[HulluJanne]

http://forums.techpowerup.com/showpost.php?p=1109854&postcount=405

This is my best on Phenom X4 9950 BE