Featured Worklog

Price Search



PC Apex Sponsor


PC Apex Sponsors



PC Apex RSS Feeds

RSS Feed for PC Apex Reviews & ArticlesRSS Feed for PC Apex PC Modding WorklogsRSS Feed for the PC Apex Daily DisturbanceRSS Feed for the latest PC Apex Site NewsRSS Feed for PC Apex Affiliate and Web NewsRSS Feed for PC Apex Deals and Steals

Go Back   Apex Community Forums // PC Apex Forums // Benchmarking // Results

Results Share your benchmark results here.

Reply
 
LinkBack Thread Tools Display Modes
Old 07-November-06, 01:23 AM   #1 (permalink)
Apex Tech Maniac
Default updated Sandra results

Well, I think that I finally (sorta) fixed my dual Opteron. The only problem now is the new mobo I got, CPU1 DIMMS1 and 2 (or some combination thereof) is dead. (It causes the system to spontaneously reboot shortly after POSTing.)

So, right now I am running the system using CPU0 DIMMS 3-4, and CPU1 DIMMS 3-4 and it seems to be okay for now. HIGHLY NOT recommended according to the Tyan website and manual, but oh well (for now). Have to RMA the board.

The good news is that it eliminates all thermal suspected problems, and dead CPU memory controllers.

It still does not explain, however, the mode of failure for the old S2875 board. I'm still suspecting an electrical fault on the AGP slot.

Anyways, I have been able to get some tests on the it, although I haven't ran the full suite of benchmarks on the system yet as it has been quite busy with work related stuff.

Here are the updated results though.



There is really no change in the MFLOPs benchmark. It is SLIGHTLY higher (by about 0.7%) possibly due to the additional 4 DIMMS and that it doesn't have to route the memory requests/IOs through CPU0 now. Still right around where I would expect it to be given that it's the same processors.

Another note is that the new board supports dual-core processors, so I MIGHT be moving to that. I don't know. I'm still very tempted to migrate to the S2915 instead.

The only downside is that I won't be able to use my Realizm 200 anymore, which means that I would have to migrate that to at least the Realizm 500 which is an additional cost. I guess that I could probably sell the system off to one of my friends who is now looking into possibly getting into the 64-bit arena.



This is where I saw the biggest improvement. Where before, I was seeing it coming up to about 86% of the theorectical maximum for PC3200 RAM (limiting at around 2.65-2.8 GB/s), I am now getting much higher than that.

That's using the Kingston KVR400D8R3AK2/1G (4x512). I would have settled with like say just a tad shy of 4 GB/s (didn't think that it would end up being THAT fast), but sure, I'd go with 10 GB/s. (That's about 21.875% short of the maximum theorectical limit.)

Not too shabby eh? That's one of the main reasons why workstation boards are what they are I suppose.

(Will be following up with the rest of the benchmark results).
Attached Thumbnails
updated Sandra results-s2885_sandra_arithmetic.jpg  updated Sandra results-s2885_sandra_memory.jpg  
alpha754293 is offline     Reply With Quote
Old 07-November-06, 06:07 AM   #2 (permalink)
Mastah Overclocker
$SOLID$ Necro's Avatar
Default Re: updated Sandra results

Is there any chance it's the PSU or Ram?

Have you tried another PSU, or let it loop Mem test?
$SOLID$ Necro is offline     Reply With Quote
Old 07-November-06, 09:41 AM   #3 (permalink)
Apex Tech Maniac
Default Re: updated Sandra results

Quote:
Originally Posted by $SOLID$ Necro
Is there any chance it's the PSU or Ram?

Have you tried another PSU, or let it loop Mem test?

No. But get this;

there are a total of 8 DIMM slots (4 per CPU).

I have my old 1 GB sticks and also my new 4x512.

Started out with the standard procedure by occupying DIMMS CPU0 DIMMA1 and 2, CPU1 DIMMA2 and 2. Failed.

I think after permutating through that for a while, I ended up occupying only CPU0DIMMA1 and CPU1DIMMA1. Failed.

Then I was reading the manual and it says that I can make it re-route the memory requests if I occupy CPU0DIMMA1 and 2 ONLY. So I did that, and got the system working, booted up, and the OS installed.

Went back to trying the original configuration; that didn't work. So went with the unconventional CPU0DIMMB1 and 2 and CPU1DIMMB1 and 2; and that worked as well.

Took about a total of 18 hours to figure that one out with the permutations trying to boot up my old install of the OS that was on the drives already, and between that and trying to figure out whether it was a memory problem or a CPU memory controller problem (which can only be caused by a overheating).

I don't think that it's a power supply or memory problem because as I've mentioned before; it doesn't make any sense why a "lower" level graphics card (VP970) would have the video signal cut out after 2 minutes, meanwhile running the higher power Realizm 200; I can run it overnight provided that the temperature was low enough???

Regardless, here are the rest of the benchmark results.



It's included here more so for completeness rather than anything else. The discussion of the significance of the scores relative to actual system performance under (almost) any scenario is long past; and it's the same discussion over and over again. Take it for what it's worth.



SPECViewperf 8.1
Quote:
Run All Summary

---------- SUM_RESULTS\3DSMAX\SUMMARY.TXT
3dsmax-03 Weighted Geometric Mean = 38.27

---------- SUM_RESULTS\CATIA\SUMMARY.TXT
catia-01 Weighted Geometric Mean = 23.94

---------- SUM_RESULTS\ENSIGHT\SUMMARY.TXT
ensight-01 Weighted Geometric Mean = 20.12

---------- SUM_RESULTS\LIGHT\SUMMARY.TXT
light-07 Weighted Geometric Mean = 16.20

---------- SUM_RESULTS\MAYA\SUMMARY.TXT
maya-01 Weighted Geometric Mean = 52.30

---------- SUM_RESULTS\PROE\SUMMARY.TXT
proe-03 Weighted Geometric Mean = 42.45

---------- SUM_RESULTS\SW\SUMMARY.TXT
sw-01 Weighted Geometric Mean = 25.14

---------- SUM_RESULTS\UGS\SUMMARY.TXT
ugs-04 Weighted Geometric Mean = 25.71

Cinebench2003
Quote:
Rendering (Single CPU): 277 CB-CPU
Rendering (Multiple CPU): 508 CB-CPU

Multiprocessor Speedup: 1.84

Shading (CINEMA 4D) : 320 CB-GFX
Shading (OpenGL Software Lighting) : 1449 CB-GFX
Shading (OpenGL Hardware Lighting) : 2762 CB-GFX

OpenGL Speedup: 8.62

Attached Thumbnails
updated Sandra results-3dmark2001se.jpg  updated Sandra results-aquamark3d.jpg  
alpha754293 is offline     Reply With Quote
Old 08-November-06, 06:22 PM   #4 (permalink)
Apex Tech Maniac
Default Re: updated Sandra results

The effect of SP2


The screenshot above is the same version of SiSoft's Sandra memory benchmark running on Windows XP Pro.

All other screenshots previous to that were running on Windows XP Pro with SP2.

As you can see, (at least with the memory benchmark); the difference is HUGE!
Attached Thumbnails
updated Sandra results-s2885_sandra_memory_nonsp2.jpg  
alpha754293 is offline     Reply With Quote
Old 08-November-06, 09:09 PM   #5 (permalink)
Apex Tech God
killert's Avatar
Default Re: updated Sandra results

Why do you post your results in this thread but have the Hwbot on another site?
Like wtf?
1. alpha754293 (Overclockers-Anonymous) - 77538.00 - Opteron 870@2.0 GHz
killert is offline     Reply With Quote
Old 08-November-06, 11:03 PM   #6 (permalink)
Mastah Overclocker
$SOLID$ Necro's Avatar
Default Re: updated Sandra results

Quote:
Originally Posted by killert
Why do you post your results in this thread but have the Hwbot on another site?
Like wtf?
1. alpha754293 (Overclockers-Anonymous) - 77538.00 - Opteron 870@2.0 GHz

ROFL..he got into a pissing match with BA a while back, so he took his toys and left...

To quote Cartman "Screw you guys, I'm going home"

He says he will repost the here "Eventually" but has not gotten around to it yet
$SOLID$ Necro is offline     Reply With Quote
Old 08-November-06, 11:06 PM   #7 (permalink)
Apex Tech God
killert's Avatar
Default Re: updated Sandra results

Quote:
Originally Posted by $SOLID$ Necro
ROFL..he got into a pissing match with BA a while back, so he took his toys and left...

To quote Cartman "Screw you guys, I'm going home"

He says he will repost the here "Eventually" but has not gotten around to it yet

Okay, That explains it.
killert is offline     Reply With Quote
Reply

Bookmarks

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is On
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Is Sandra biast? St8ic Results 6 07-May-05 10:31 AM
Sandra CPU Scores mot19 Results 26 03-January-05 12:30 AM
Sandra 2004 Scores godfoot Apps 32 30-August-04 08:03 PM
Using Sandra to help tweak your vcore Shadowman0000 Intel CPU/Motherboard OC 1 09-January-04 03:37 PM


All times are GMT -5. The time now is 11:57 PM.


Powered by vBulletin® Version 3.8.3
Copyright ©2000 - 2009, Jelsoft Enterprises Ltd.
Search Engine Optimization by vBSEO 3.3.0
Copyright PCApex.com, GameApex.com, ForumApex.com 2001 - 2008
Advertisements

Page generated in 0.20800 seconds with 10 queries