Rundown to SuperPrime 1.0Final : Need your help!

Printable View

Show 100 post(s) from this thread on one page

12-06-2007, 02:13 PM
Ad1tya

SneakPeak at 0.9.8.1..

Much cleaner layout.. Also, a new feature coming up..

http://forums.erodov.com/imagehostin...872b7cd5cf.png

Non-Optimised run.. =X
12-07-2007, 12:25 AM
KTE

Well here's 2.7GHz Phenom 9500 for your comparison ;)

http://img151.imageshack.us/img151/969/supersy7.png
12-07-2007, 12:45 AM
KiD0M4N

Quote:

Originally Posted by KTE

Well here's 2.7GHz Phenom 9500 for your comparison ;)

http://img151.imageshack.us/img151/969/supersy7.png

Glad to have you back on the thread mate.

0.9.8.2 version is out, and I took your advise. Pls give it a try and see if the detection of system parameters is improved.

http://forums.erodov.com/local_links...tid=7#linkid17

PS: KTE, didn't you get any auto update messages from 0.9.7?
12-07-2007, 01:19 AM
KTE

Nope, I didn't get any update reminders. But it runs the check with the latest version fine.

Congrats on the new version! I like it, the layout is nicer, runs smooth, more CPU usage (92-98%), brilliant detection. It also runs faster. :toast:

Here's a quick run with it:
http://img84.imageshack.us/img84/2654/superil9.th.png
12-07-2007, 01:28 AM
KiD0M4N

Many thanks to Franck/CPUz :) :)

KTE, need one help from you. Since you are the only guy with a L3 cache here, care to PM/email me a register dump from your system?

Download CPUz 1.42 and About > Register Dump.
You already have my email id.

Will be able to fix that small little nagging bug once and for all.

I think the reduced CPU utilization compared to C2D, C2Q and Clovertowns is because of the TLB workaround? I think the stress on the TLB is causing the CPU to slow down substantially.

Anyone with a C2D care to post a result at around 2.6G? For a comparo against this Phenom.

Quote:

Originally Posted by KTE

Nope, I didn't get any update reminders. But it runs the check with the latest version fine.

Congrats on the new version! I like it, the layout is nicer, runs smooth, more CPU usage (92-98%), brilliant detection. It also runs faster. :toast:

Here's a quick run with it:
http://img84.imageshack.us/img84/2654/superil9.th.png
12-07-2007, 01:43 AM
KTE

Quote:

Originally Posted by KiD0M4N

Many thanks to Franck/CPUz :) :)

KTE, need one help from you. Since you are the only guy with a L3 cache here, care to PM/email me a register dump from your system?

OK will do. I'm in work but I'm off because of two broken knuckles. :D

Quote:

I think the reduced CPU utilization compared to C2D, C2Q and Clovertowns is because of the TLB workaround? I think the stress on the TLB is causing the CPU to slow down substantially.

I don't experience the bug unless I go above 2.16GHz NB or above 2.86GHz CPU so I never applied the patch. ;)

Quote:

Anyone with a C2D care to post a result at around 2.6G? For a comparo against this Phenom.

I have one, I'll match the timings and post it. Gimme 30-45 minutes, I'm one handed. :D
12-07-2007, 02:09 AM
KiD0M4N

Quote:

Originally Posted by KTE

OK will do. I'm in work but I'm off because of two broken knuckles. :D
I don't experience the bug unless I go above 2.16GHz NB or above 2.86GHz CPU so I never applied the patch. ;)

I have one, I'll match the timings and post it. Gimme 30-45 minutes, I'm one handed. :D

Wah... thats cool. Thanks bhaijan :P

I am out of sorts seriously... why isnt the utilization pegged to 100% like in C2Ds... am I missing something somewhere.

Waiting for the register dump. :up:
12-07-2007, 03:01 AM
KTE

1 Attachment(s)

Here its is. Off to start C2 system now.
12-07-2007, 03:06 AM
KiD0M4N

Quote:

Originally Posted by KTE

Here its is. Off to start C2 system now.

THANKS and thanks....

Registers :P yum yum....
12-07-2007, 04:00 AM
tiborrr

Can you integrate some sort of encripted validation in order to keep validate result? If you can, i will certainly open an competition ladder for this application on our community.
12-07-2007, 04:06 AM
KTE

Here's an E6750 on P35 running 365x7 2555MHz (close enough?) with DDR2-912 4-4-4-4 2T PL5 (closest I could do): 58.62s ;)

http://img149.imageshack.us/img149/4875/superqz4.th.png

I'm going to run the same test again on Phenom when I get time using only two cores and see what the difference from 2 core vs 4 core is.
12-07-2007, 04:21 AM
Mindmaster

C2D 6300 @458 MHz FSB
The Frequencies and the Multi are wrong

http://www.abload.de/thumb/superprimetd0.jpg
12-07-2007, 04:36 AM
KiD0M4N

Quote:

Originally Posted by Mindmaster

C2D 6300 @458 MHz FSB
The Frequencies and the Multi are wrong

http://www.abload.de/thumb/superprimetd0.jpg

Hey,

First of all thnx for trying this out.

I have a request. Care to save the result (Right click on the result display > Save to file...) and host it here? I would love to figure out why its giving such a weird result.

Regards,
Karan Misra
12-07-2007, 04:56 AM
KTE

Ask Mindmaster for CPU-Z screenshots. Then you can compare if its just your app or others too. ;)
12-07-2007, 04:58 AM
KiD0M4N

Quote:

Originally Posted by KTE

Here's an E6750 on P35 running 365x7 2555MHz (close enough?) with DDR2-912 4-4-4-4 2T PL5 (closest I could do): 58.62s ;)

http://img149.imageshack.us/img149/4875/superqz4.th.png

I'm going to run the same test again on Phenom when I get time using only two cores and see what the difference from 2 core vs 4 core is.

Yup, that will shed some light.
12-07-2007, 07:02 AM
KTE

Here ya go.

X2 5000+ BE at same mult/speed/HT (etc) as Phenom 9500 (2 threads). Couldn't keep RAM frequency the same but it shouldn't matter that much. ;)

X2 - Phenom
http://img142.imageshack.us/img142/6263/supertm1.th.png - http://img487.imageshack.us/img487/6...rk10dq6.th.png

This is what I don't understand about the bench. How can 2 threads on 2 cores (above) get the same or higher time than 4 threads on 2 cores (below)? :confused:
http://img490.imageshack.us/img490/7...2corjl1.th.png
12-08-2007, 01:53 AM
G H Z

Looks good, I noticed it doesn't run on 64-bit OS.
12-08-2007, 01:56 AM
Movieman

Quote:

Originally Posted by G H Z

Looks good, I noticed it doesn't run on 64-bit OS.

Yup, it does. This was on win2K3 ent 64 bit:
http://www.xtremesystems.org/forums/...6&postcount=25
12-08-2007, 02:29 AM
KiD0M4N

Quote:

Originally Posted by G H Z

Looks good, I noticed it doesn't run on 64-bit OS.

Quote:

Originally Posted by Movieman

Yup, it does. This was on win2K3 ent 64 bit:
http://www.xtremesystems.org/forums/...6&postcount=25

Thanks for the confirmation.

I have personally run it on Vista X64 and tested it. Ran perfectly. G H Z did you try the latest version? 0.9.8.2

Let me know.

Regards,
Karan
12-08-2007, 04:04 PM
G H Z

A re-boot fixed the initial problem where sprime would simply dissapear after execution of the bench - something about checking single threaded results or similar in log window (.NET 3.5), now I can complete the bench.

Don't mind the temp differences, one quad is B3, the other G0.

http://img.photobucket.com/albums/v3...y/sprime_2.jpg

I do get a system error on completion, could be that this is a server with more security than a desktop.

http://img.photobucket.com/albums/v3...prime-er_2.jpg

Either way this is a nice program, only changes I'd suggest are to consider a a different naming scheme for the different test sequences, test/defense against speed hack programs & online validation.

Great work so far :)
12-09-2007, 04:48 AM
KiD0M4N

Thanks so much G H Z :up: :up: :up: :up: :welcome: :woot:

I am working feveresly on the online validation and result submission. If you see, I have worked some kind of encryption into the results, and I will always be around (unless Global warming floods my home, i.e.) to take care of "those dem haxxers!" if need be. A method of checking is already incorporated (something to do with the single threaded checking :P)

I will definitely work on the naming scheme of the runs. Any suggestions?

I am myself developing this on VS2008 (.NET 3.5) but the application profile set to .NET 2.0 so as to make sure it runs everywhere with atleast 2.0 installed.

I will definitely try this on a server enviroment (VMware, here I come) and try to get that error fixed.

@David: Did you get the same error in the ~19 sec 1E8 run of yours? I think both are on Windows Server 2003 OS? (both yours and G H Z's run i.e.)

Regards,
Karan

Quote:

Originally Posted by G H Z

A re-boot fixed the initial problem where sprime would simply dissapear after execution of the bench - something about checking single threaded results or similar in log window (.NET 3.5), now I can complete the bench.

Don't mind the temp differences, one quad is B3, the other G0.

http://img.photobucket.com/albums/v3...y/sprime_2.jpg

I do get a system error on completion, could be that this is a server with more security than a desktop.

http://img.photobucket.com/albums/v3...prime-er_2.jpg

Either way this is a nice program, only changes I'd suggest are to consider a a different naming scheme for the different test sequences, test/defense against speed hack programs & online validation.

Great work so far :)
12-09-2007, 07:15 AM
Movieman

Quote:

Originally Posted by KiD0M4N

Thanks so much G H Z :up: :up: :up: :up: :welcome: :woot:

I am working feveresly on the online validation and result submission. If you see, I have worked some kind of encryption into the results, and I will always be around (unless Global warming floods my home, i.e.) to take care of "those dem haxxers!" if need be. A method of checking is already incorporated (something to do with the single threaded checking :P)

I will definitely work on the naming scheme of the runs. Any suggestions?

I am myself developing this on VS2008 (.NET 3.5) but the application profile set to .NET 2.0 so as to make sure it runs everywhere with atleast 2.0 installed.

I will definitely try this on a server enviroment (VMware, here I come) and try to get that error fixed.

@David: Did you get the same error in the ~19 sec 1E8 run of yours? I think both are on Windows Server 2003 OS? (both yours and G H Z's run i.e.)

Regards,
Karan

I didn't get any errors at all. Just installed, bumped the machine up a little and let it fly..
Perfect.:D
12-09-2007, 09:17 AM
FrogBite

1 Attachment(s)

here's mine, the cpu speed is a bit wrong...
12-09-2007, 09:19 AM
Movieman

Quote:

Originally Posted by FrogBite

here's mine, the cpu speed is a bit wrong...

I hope so or you just cut my clovers time in 1/2:rofl:
12-09-2007, 10:43 AM
KiD0M4N

Quote:

Originally Posted by FrogBite

here's mine, the cpu speed is a bit wrong...

Quote:

Originally Posted by Movieman

I hope so or you just cut my clovers time in 1/2:rofl:

Na, that was a 5E7 run... + not valid (see the brown line below the score) + I have no clue why it didn't detect your hardware properly.

FrogBite, care to send me your CPUz dump so I can fix it right up? Thanks again.

Show 100 post(s) from this thread on one page