CrystalMark 2004 Competition! Stress your System!

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Zebo

Elite Member
Jul 29, 2001
39,398
19
81
Originally posted by: stevty2889
34,698

Case in point. These overclocked San deigos at 2800Mhz will destory a 3.8 Prescott in every real world bench there is. Not here, Not with synthetics. They loose by over 10%!!
 

Zebo

Elite Member
Jul 29, 2001
39,398
19
81
Originally posted by: Tool
30,924

i am curious why my FPU blows so hard...

Well your mem and alu is whoppin me and we have exact same ram same timings too..

some bios setting is hurting you.

I'm sorry I can't help cause I don't own board but inside my board a setting called "maximum performance" whateverthehell that means increased FPU by a lot....try these links to find optimum settings inside DFI.

nF4 step-by-step setup: http://www.dfi-street.com/forum/showpost.php?p=65304&postcount=49
RGone's DFI nF4 BIOS guide: http://www.dfi-street.com/forum/showthread.php?t=5866
 

Tool

Senior member
Feb 5, 2005
784
0
71
yeah....i'm a lurker on dfi street. i've been testing each individual setting....ugh. Guess i need to keep going. :yawn:
 

MADMAX23

Senior member
Apr 22, 2005
527
0
0
Just posted the Marks for now! Check them out in page 1!
I would suggest you to complete your own specs if any is missing.
I will keep it up to date.
 

GuitarDaddy

Lifer
Nov 9, 2004
11,465
1
0
Originally posted by: Zebo
Originally posted by: stevty2889
34,698

Case in point. These overclocked San deigos at 2800Mhz will destory a 3.8 Prescott in every real world bench there is. Not here, Not with synthetics. They loose by over 10%!!

Yeah, it definately rewards HT! Is there anything that proves HT has a real world performance advantage? I,m not even sure what kind of apps are optimized for HT?
 

stevty2889

Diamond Member
Dec 13, 2003
7,036
8
81
Originally posted by: GuitarDaddy
Originally posted by: Zebo
Originally posted by: stevty2889
34,698

Case in point. These overclocked San deigos at 2800Mhz will destory a 3.8 Prescott in every real world bench there is. Not here, Not with synthetics. They loose by over 10%!!

Yeah, it definately rewards HT! Is there anything that proves HT has a real world performance advantage? I,m not even sure what kind of apps are optimized for HT?

I don't think it's multithreaded, I'll run it again with HT turned off and see if it makes a noticable differance..

EDIT: With HT off my score dropped to 26849.
NO HT......................HT
26849.....................34698
ALU 8624................12414
FPU 9208.................13313
MEM 9017................8971
 

GuitarDaddy

Lifer
Nov 9, 2004
11,465
1
0
Wow HT increases your score by 29%! I really wonder which real world apps can take advantage of HT? Video editing/encoding? any others
 

Hurricane Andrew

Golden Member
Nov 28, 2004
1,613
0
76
Mark 28265
ALU 5075
FPU 6087
MEM 3715
HDD 4650
GDI 3448
D2D 2832
OGL 2458

This was on my work PC:
CPU Name : Intel Pentium 4 (Northwood)
Vendor String : GenuineIntel
Name String : Intel(R) Pentium(R) 4 CPU 2.40GHz
Architecture : x86
Process Rule : 0.13 um
Platform : Socket 478 [2]

Fortunately, Newegg is on top of things, and I should be receiving my new work PC parts tomorrow: an Athlon 64 3400+. Should be a nice bump up in performance :D
 

Shimmishim

Elite Member
Feb 19, 2001
7,504
0
76
mine

34,044
ALU = 10351
FPU = 12771
MEM = 10922

Clawhammer 3500+ @ 2.81 ghz
OCZ VX pc3200 @ 1.5-2-2-6 1:1 running 256 HTT

not bad for .13 tech chip :)
 

MADMAX23

Senior member
Apr 22, 2005
527
0
0
Originally posted by: Shimmishim
mine

34,044
ALU = 10351
FPU = 12771
MEM = 10922

Clawhammer 3500+ @ 2.81 ghz
OCZ VX pc3200 @ 1.5-2-2-6 1:1 running 256 HTT

not bad for .13 tech chip :)

WOW, very nice results. Ram 1T?

 

GuitarDaddy

Lifer
Nov 9, 2004
11,465
1
0
Nice score Shim! I think this really shows that this program loves memory bandwidth.
Your score beats my SD3700+ by 1400 points, even though I was running 2860mhz (50mhz higher).

But your ram at 256mhz 1.5-2-2-6 clobbers my ram at 236 2-3-2-7, this shows up in the MEM score as well as the FPU score.

The more scores I see, the more I realize this program is heavily biased towards Intel. It heavily rewards HT which AMD's don't have, and it heavily rewards memory bandwidth, which has been proven to have little performance affect on A64's. And it seems to place very little value on highend graphics cards, so if your primary performance requirements are gaming and normal single threaded apps this benchmark is worthless.
 

MADMAX23

Senior member
Apr 22, 2005
527
0
0
Yeah, You are right GuitarDaddy, this program is optimized for Intel Processors.However, It is challenging for us, AMD users, to run it in order to see how our different settings affect the scores.
I wouldn´t compare AMD scores to Intel's GuitarDaddy since Intel cpus win in most of synthetic benchmarks.
GD,What do you think about 1010.001 Bios?
What maximum "safe" voltage for a Winchester?
 

Shimmishim

Elite Member
Feb 19, 2001
7,504
0
76
Originally posted by: GuitarDaddy
Nice score Shim! I think this really shows that this program loves memory bandwidth.
Your score beats my SD3700+ by 1400 points, even though I was running 2860mhz (50mhz higher).

But your ram at 256mhz 1.5-2-2-6 clobbers my ram at 236 2-3-2-7, this shows up in the MEM score as well as the FPU score.

The more scores I see, the more I realize this program is heavily biased towards Intel. It heavily rewards HT which AMD's don't have, and it heavily rewards memory bandwidth, which has been proven to have little performance affect on A64's. And it seems to place very little value on highend graphics cards, so if your primary performance requirements are gaming and normal single threaded apps this benchmark is worthless.

@madmax, yes i used 1T for that run. and as the max for winchester voltage on air... i'd say 1.65 but most would probably say 1.6. i say watch your temps as you increase voltage

yeah... my memory is pretty tweaked.

i changed my trtw value from 2 to 1 and got a pretty big increase.

my first run i got a lil less than 32k. changed my trtw to 1, and from 10 x 280 with the 9/10 divider to 11 x 256 no divider and got 34k+.

i have to agree that it is sensitive to mem bandwidth and maybe a reason why this bench does heavily favor intel chips.
 

Zebo

Elite Member
Jul 29, 2001
39,398
19
81
Max - My ram is 1GB Ballistix @ 241 Mhz 2.5-2-2-7 1T

Hey DP you're gonna double that score in two weeks:D howz the manual labor?:p
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
32,142
32,717
146
Originally posted by: Zebo
Max - My ram is 1GB Ballistix @ 241 Mhz 2.5-2-2-7 1T

Hey DP you're gonna double that score in two weeks:D howz the manual labor?:p
I am like a piece of iron now :D Using a 10-12 kilo break bar to bust up cement for 5-6hrs@ a time is a awesome workout braddah, the good bank is just the bonus. We are building operations support building #2; I was about 100 yrds away from the shuttle being taken into the VAB. Unbelievable tech going in, 42"plasmas all over the building, and you should see the communications room :Q We call the plasma rigs the puppet playhouses :p

The Operations Support Building II at Kennedy Space Center is designed to exceed EO13123 and 10 CFR 435 energy efficiency requirements and will use at least 47 percent less energy than the temporary, modular offices it will replace. The project will incorporate an automatic irrigation system, 100 percent native plants with low water requirements, high efficiency lighting with occupancy sensors anddaylight-compensating dimmer controls, variable air volume HVAC systems with variable frequencydrives on air handlers and chilled water pumps, and advanced filtration to maintain adequate indoor air quality

Of course I will be buying a X2 to play with :beer: I'll stick to air cooling though so you should beat my overclock.