New Athlon 64 System Benchmarks...I might need some help!

[jdogg707]

Just got it together today...

System:

AMD Athlon 64 3200+ 2.0GHZ CO Stepping (10 x 200MHZ) (AMD Driver Version 1.1)
MSI K8T Neo-FISR2 Motherboard (BIOS 1.1)
1GB Geil Ultra Platinum PC3500 DDR RAM (2.5-3-3-6)
ATI Radeon 9600XT 128MB AGP Video Card (Default Clock w/Overdrive Enabled) (ATI Catalyst 4.1)
Sound Blaster Audigy 2 ZS Platinum
Western Digital 74GB 10K RPM SATA Raptor Hard Drive (SATA 0 on Via VT8237 Southbridge)
Hitachi 160GB 7200RPM SATA Hard Drive (SATA 1 on Via VT8237 Southbridge)
NEC ND-2500A 8X DVD+/-RW
Logitech Cordless MX Duo
Samsung 181T 18.1" LCD
Windows XP Professional w/SP1 (Fully Updated)
Zalman CNPS7000A-Cu on Full (AS5 Applied to CPU and NorthBridge)
Antec SLK3700-BQE w/350W Antec PSU (1 x 120mm Antec Fan Blowing Out, 1 x 120mm Enermax Variable Speed Fan for Intake)

Temperature:

CPU - 34C Idle / 49C Load (SETI for 2 Hours)
Case - 30C Idle / 34C Load (SETI for 2 Hours)
GPU - 30C Idle / 45C Load (Gaming for 2 Hours)

Benchmarks:

3DMark 2001 SE:

Total - 14023

3DMark 2003:

Total - 3923
CPU Score - 633

AquaMark3:

Total - 32,348
GPU Score - 3,963
CPU Score - 8,802

PCMark 04:

Total - 3935
CPU Score - 3708
Memory Score - 3521
Graphics Score - 2854

PCMark 2002:

CPU Score - 6441
Memory Score - 8521
Hard Drve Score - 1633

SiSoft Sandra 2004:

CPU Arithmetic -
Dhrystone ALU 8330 MIPS
Whetstone FPU/iSSE2 3159/4095

CPU Multi-Media -
Integer iSSE2 14942 it/s
Float iSSE2 19710 it/s

Memory Bandwidth -
INT Buffered iSSE2 2985MB/s
Float Buffered iSSE2 2981MB/s

File system Benchmark -
Drive index 45542kb/s

AIDA32 v. 3.9:

Memory Read - 2925MB/s
Memory Write - 1223MB/s

Unreal Tournament 2003 Demo v2206:

DM-Antalus Flyby - 135.92FPS
DM-Antalus Botmatch - 87.27FPS

DM-Asbestos Flyby - 233.75FPS
DM-Asbestos Botmatch - 95.73FPS

DM-Citadel Flyby - 144.74FPS
DM-Citadel Botmatch - 72.60FPS

Prime95 Benchmark:

384K - 23.606ms
448K - 28.335ms
512K - 32.051ms
640K - 40.249ms
768K - 49.399ms
896K - 59.574ms
1024K - 67.298ms
1280K - 90.187ms
1536K - 110.255ms
1792K - 133.409ms
2048K - 149.082ms


Notes:

A Fresh copy of Windows XP Professional was installed, all drivers were updated to the newest available releases.


Comments:

Some of these scores seem fairly low from what I have seen of the Athlon64. My 3Dmark03 CPu Score and PCMark 2002 CPU Score especially. My temps also seemed a little high for my the CPU cooler I have, but I believe that it is the low operation of the Case fans that bring the temps up. Anyone else have numbers like these or is there something wrong with my system? My 9800NP pooped out on me the other day, so I am using the 9600XT until I get my replacement sometime next week...so 3D scores I'm not terribly worried about. If anything I would blame it on the motherboard, but would like some other opinions.

Thanks!

 

[Markfw]

Your case temp is a little high for the winter, but maybe its hot in North Carolina and you don't have AC, but those are fine temps. Mine is 31c/42c right now under seti for 2 hours or more.

As far as the benches, I don't really look at those that much. The real question is, how does the computer seem to you, and how do games play compared to your old one ??

Some feedback here !!!!

 

[Dman877]

Most of the benches look like its the 9600 holding em back. As for the other stuff, the A64 doesn't seem to do well in PCMark. Like the other guy said, worry less about becnhies, if the system feels fast don't get hung up on it. Temps do seem a bit high, my 3000 never goes over 33 or so with a venus 12 on it.

 

[jdogg707]

The system feels more responsive, but I am mainly attributing that to the 74GB Raptor I am now using. I can't really compare 3D benches or how it feels, because of course it is going to be slower, sinceI don't have the same card. I don't put a lot of faith in benches, but they do serve a purpose, by letting you know if your system is performing at it's optimum level. If I spend a bunch of money on a computer, I want it to perform the best that it can, whether I can tell the difference or not.

 

[Markfw]

OK, I can buy that (until you get your 9800np back), so how long does the command line version of seti take to do the benchmark unit and/or the average of the last 5 seti packets you did ?? My 3200+ is still at 2 1/2 hours, and this a64 does them in 1:45 to 2:15, with the average right at 2 hours. That is pretty much just CPU if you are running the comand line version. Most other things depend of the video card or HD a lot. I am sure there are some others, this one is the only one that comes to mind right now.

 

[jdogg707]

Alright, how do I set it up to Benchmark with SETI? I have the command line version but use SETI Driver to see my results. Also, is there anyway you could run 3DMark03 CPU Scores @ Default on your 3000+? I have heard about the PCMark problems, but my 3DMark03 CPU score looks about 200 points lower than every review/benchmark comparison I have seen.

 

[DAPUNISHER]

I have a bunch of benchies here you can check out and your CPU, memory, and HDD scores are all quite good so stop sweating it Did you run the [ H ] UT bench? If so try just running the default so you can see how the Bot match score compares to the ones I posted@higher clockspeeds. As was stated the vid card is the only thing bonking.

 

[DAPUNISHER]

Your '03 CPU score is low but I'd pay that no mind since 3DMock sux

 

[RussianSensation]

Originally posted by: jdogg707[/i]
Just got it together today...

Prime95 Benchmark:

384K - 23.606ms
448K - 28.335ms
512K - 32.051ms
640K - 40.249ms
768K - 49.399ms
896K - 59.574ms
1024K - 67.298ms
1280K - 90.187ms
1536K - 110.255ms
1792K - 133.409ms
2048K - 149.082ms

I have 2.6@3.2 and most of my scores are slightly faster than yours, but I know Athlon 64 sucks at synthetic benches and they dont matter much at all....because it comes up great in games and real world applications....Either way your 14000 in 3dmark is a bit too low, you should be hitting 15-16 thousand. Check if you are using all performance settings, AA /AF disabled, TruForm Disabled, VSYnc off, etc.

What surprised me the most is the Prime 95 benchmark because I would imagine Athlon 64 is a lot faster at Arithmetic caclulations and has lower latency. I have same PC3500 UP Geil ram as you and I get:

384K - 11.179ms
448K - 13.302ms
512K - 15.135ms
640K - 18.198ms
768K - 22.084ms
896K - 26.160ms
1024K - 29.346ms
1280K - 38.641ms
1536K - 47.513ms
1792K - 56.403ms
2048K - 63.976ms

That's more than twice as fast. I am not trying to brag, but can someone try to explain to me why the p4 does so much better in this benchmark in particular? Also, that Geil Ram is rated at 2-3-3-6 @ 433mhz so you can run 10x200 at 2-3-3-6 at least. If you need reassurance and stability bump the voltage to 2.8. Mine does 2-2-2-6 at 188mhz (376mhz effective) so I dont see the need for you using 2.5 latency. I know the improvement might be 0-2% but every % helps little by little.

About the temperatures I dont think if you are running below the maximum threshhold that it matter much. I mean I highly doubt that my p4 will last longer if i run it at -30 or 65*C. Temperatures matter a lot when they interfere with the maximum overclocking potential, other than that don't worry too much.

nice rig by the way sorry to hear about your 9800np

my 3.2 with Radeon 8500 in Unreal Tournament 2003:

at 1024x768 bench max settings no sound

Flyby: 147.54fps
Botmatch: 77.24

How did you get 6 numbers for your benchmark???? hmmm.for me after it ran it, it just popped 1 window.

 

[jdogg707]

Originally posted by: DAPUNISHER
I have a bunch of benchies here you can check out and your CPU, memory, and HDD scores are all quite good so stop sweating it Did you run the [ H ] UT bench? If so try just running the default so you can see how the Bot match score compares to the ones I posted@higher clockspeeds. As was stated the vid card is bonking the only thing bonking.

I downloaded the demo off of IGN.com and just ran the benchmarks in the Benchamrking folder @ 1024x768. Is there a difference in that and the [ H ] UT bench? If so where do I get it/set it up?

Thanks for all the quick replies, I am just looking to make sure there isn't something that I am missing in order to get the best performance.

 

[Farmer]

The RAM timings may also contribute, as you are running at 2.5-3-3. If your RAM is rated for it, run it at CAS 2. It might help a little (though you may want to stay at 2.5 for stability's sake).

As Markfw900 would know, my case temp and CPU temp are much hotter than yours, so there is nothing to worry about (I think mine is around 45c case/58c CPU on max load on a Tt Venus 12 with AS5). Your HSF is of wonderful quality, as is your thermal compound. I also have the MSI board, I doubt the manufacturer as much to do with performance (most K8T800 mobos perform like eachother).

My benchmark/games suite isn't nearly as extensive as yours (I havn't bothered to acquire 3Dmark, though I have really wanted to), but I did download Sandra, and here are my scores, and my specs:

A64 3200+ @ 2000.0MHz
1GB (512x2) PC3200 2-3-2
Radeon 9800XT @ stock speeds
Hitachi 160GB

SiSoft Sandra Win32 x86 2004.10.9.89

CPU Arithmetic -
Dhrystone ALU :8343 MIPS
Whetstone FPU/iSSE2: 3149/4084 MFLOPS

CPU Multi-Media -
Integer iSSE2: 14915 it/s
Float iSSE2: 19650 it/s

No surprise here, we have the same CPU clocked at the same speed. There are slight differences, but mostly because of test error (If you run your test multiple times, you get an average, I won't bother doing that). The A64 isn't the multimedia king like the P4HT, but who cares?

Memory Bandwidth -
INT Buffered iSSE2: 3010MB/s
Float Buffered iSSE2: 3011MB/s

I run Crucial XMS3200LLPT 400 2-3-2. Again, no real difference. Same single channel DDR, same on-CPU memory controller, same motherboard. I am running at CAS 2, however. Again, single channel gets blown out of the water by dual channel for pure MB/s.

File system Benchmark -
Drive index 34604kb/s

Im assuming this is your Raptor. Your drive blows mine away (I have the Hitachi SATA 160GB you use as a secondary as my primary).

Thanks for this thread. I've always wanted to compare benches to make sure I didn't screw something up. But, with the A64, like RussianSensation and Markfw900, computer performance cannot be determined solely through synthetic benchmarks. You know this because you know your A64 is fast.

 

[jdogg707]

Originally posted by: Farmer
The RAM timings may also contribute, as you are running at 2.5-3-3. If your RAM is rated for it, run it at CAS 2. It might help a little (though you may want to stay at 2.5 for stability's sake).

As Markfw900 would know, my case temp and CPU temp are much hotter than yours, so there is nothing to worry about (I think mine is around 45c case/58c CPU on max load on a Tt Venus 12 with AS5). Your HSF is of wonderful quality, as is your thermal compound. I also have the MSI board, I doubt the manufacturer as much to do with performance (most K8T800 mobos perform like eachother).

My benchmark/games suite isn't nearly as extensive as yours (I havn't bothered to acquire 3Dmark, though I have really wanted to), but I did download Sandra, and here are my scores, and my specs:

A64 3200+ @ 2000.0MHz
1GB (512x2) PC3200 2-3-2
Radeon 9800XT @ stock speeds
Hitachi 160GB

SiSoft Sandra Win32 x86 2004.10.9.89

CPU Arithmetic -
Dhrystone ALU :8343 MIPS
Whetstone FPU/iSSE2: 3149/4084 MFLOPS

CPU Multi-Media -
Integer iSSE2: 14915 it/s
Float iSSE2: 19650 it/s

No surprise here, we have the same CPU clocked at the same speed. There are slight differences, but mostly because of test error (If you run your test multiple times, you get an average, I won't bother doing that). The A64 isn't the multimedia king like the P4HT, but who cares?

Memory Bandwidth -
INT Buffered iSSE2: 3010MB/s
Float Buffered iSSE2: 3011MB/s

I run Crucial XMS3200LLPT 400 2-3-2. Again, no real difference. Same single channel DDR, same on-CPU memory controller, same motherboard. I am running at CAS 2, however. Again, single channel gets blown out of the water by dual channel for pure MB/s.

File system Benchmark -
Drive index 34604kb/s

Im assuming this is your Raptor. Your drive blows mine away (I have the Hitachi SATA 160GB you use as a secondary as my primary).

Thanks for this thread. I've always wanted to compare benches to make sure I didn't screw something up.

Wow, those benches do look very similar...so maybe it's just a mething. I'm going to go see how low I can get my RAM to run and also play with some overclocking...

 

[jdogg707]

Ok, I am now running at 2.16GHZ, my RAM is at 2-3-3-6 (2.8V). My scores look better, but Sandra reports myPCI bus is running at 48MHZ...that can't be good! Either way, I'm going to run Prime and see how it goes...be back in a few.

 

[RussianSensation]

are you sure its not 38mhz? it shouln't run at 66/33 because VIA lacks AGP/PCI but it shouldnt be 48 either. Check that for sure.

 

[mechBgon]

I have an A64 300+ on the way and I'm going to do some benchmarks with my stock-clocked 2500+ / nForce2 Ultra400 rig before & after I swap. Planned benchies:

• The infamous WinZip/UT benchie
• Full, no-holds-barred antivirus scans of nasty multi-layered compressed files containing more compressed files... the other day I checked the AV log on my system and the number of files it unpacked and scanned during its daily AV scan was over 800,000 files :Q
• Time to re-launch a slow-launching application direct from RAM after it's gotten cached in RAM (PhotoImpact 6)
• Time to decompress a ~540MB self-extracting .EXE that contains my Office2000 Disc 2 AIP
• Time to compress the ~20,000 files that comprise my Disc1 and Disc2 AIPs, totalling nearly 1GB of data
• Adaptec SCSIBench, for testing the PCI throughput from my SCSI cards
• SmileMark 2004 and Seat-Of-The-Pants-Mark 2004, which counts as much as anything else when you work with a tool 4 to 6 hours a day

Should be fun!

 

[microAmp]

Originally posted by: mechBgon
I have an A64 300+ on the way....

That's a pretty slow processor!

For the OP:

My temps are pretty much like yours.

3DMark01 I got 18200 with 9700 Pro @ 2.1 GHz

SiSoft Sandra 2004:

CPU Arithmetic -
Dhrystone ALU 8766 MIPS
Whetstone FPU/iSSE2 3310/4299

CPU Multi-Media -
Integer iSSE2 15673 it/s
Float iSSE2 20680 it/s

Memory Bandwidth -
INT Buffered iSSE2 3083 MB/s
Float Buffered iSSE2 3083 MB/s

Bah, Cool & Quite were giving me benchmarking problems.

 

[jdogg707]

Speaking of 380...it's weird, Windows XP reports my CPU speed as being 380MHZ instead of 2.16GHZ....Prime is still going strong...mechBgon...will look forward to seeing your benches once you get the chip...

 

[imported_Ziggy]

Originally posted by: jdogg707
Speaking of 380...it's weird, Windows XP reports my CPU speed as being 380MHZ instead of 2.16GHZ....Prime is still going strong...mechBgon...will look forward to seeing your benches once you get the chip...

Doesn't athlon 64 have some type of cpu scaling where it keeps temps down by not using the max processor power unless you need it?

 

[jdogg707]

Yes, it does have a CPU scaling feature called "Quiet n' Cool" but I don't have it enabled.

 

[jdogg707]

Stable at 2.16...going for 2.2 now

 

[axelfox]

those numbers are no good. Send the chip to me

 

[RussianSensation]

I just read this:

Top Processor Summary Benchmarks

I think it's pretty clear what the best processor is today.

I'll quote: "However, there are two types of tasks where Athlon 64 is no rival to Intel?s solutions: audio/video encoding and Photoshop. If it were not for these tasks, we would have every right to call Athlon 64 the best CPU in the today?s market."

As you can tell the % differences between Athlon 64 and P4 are far greater than between p4 and XP. ...And everyone pretty much agreed the XP couldn't compete with p4 at the end of its line, so why is everyone having such a hard time deciding between p4 and athlon 64? Apart from the advantages mentioned in the quote above, 64 is simply a better solution.

 

[jdogg707]

I agree completely, the only problem with the Athlon 64 right now is the fact that there is no agreed best platform to run it on and the upcoming move to Socket 939. I bought my stuff with the hopes of running it until next September-December and doing a mjor upgrade depending on what's on then.