Switched to C2D...

[yh125d]

My old system

A64 4000+ @ 2.6gHz
1.5 gigs ddr333 ram
7600gt

old score - 4002


My new system

e4300 @ 1.8
1 gig ddr2-667
7600gt

new score - 3227

http://service.futuremark.com/orb/resul....jsp?projectType=14&XLID=0&UID=8971795


What could be causing this?

 

[Atty]

Loss in ram?

I haven't looked at E4300 benchmarks but could it be it wasn't that much of an improvement? Try overclocking it, then run the test.

 

[yh125d]

I have a ram usage meter goin, and it never got over 900mb used

 

[LittleNemoNES]

less ram and slower overall processor.

 

[yh125d]

Are there oatches or versions on 3dmark that are multithreaded? that's where I'd notice a difference I guess

 

[Jeff7181]

While a 2.6 GHz Core 2 Duo is faster than a 2.6 GHz Athlon 64, expecting the Core 2 architecture to make up for an 800 MHz clock speed advantage is asking a lot.

 

[yh125d]

wtf?? I ran it again and got this

http://service.futuremark.com/orb/resul....jsp?projectType=14&XLID=0&UID=8972113

4022

 

[postmortemIA]

Originally posted by: Jeff7181
While a 2.6 GHz Core 2 Duo is faster than a 2.6 GHz Athlon 64, expecting the Core 2 architecture to make up for an 800 MHz clock speed advantage is asking a lot.

Yeah, A64 has 44% higher clock, and definitely C2D is not 44% faster per clock than A64. That would be insane... unless we replace A64 with P4

 

[imported_K3N]

Overclock the e4300 to 3.0ghz and come back with the results. Dont expect to get much on stock speeds with that processor.

 

[Cheezeit]

Originally posted by: yh125d
wtf?? I ran it again and got this

http://service.futuremark.com/orb/resul....jsp?projectType=14&XLID=0&UID=8972113

4022

Did you run the test after a reboot and get a higher score? My 3dmark score doubles after a reboot.

 

[yh125d]

no, i just ran it again, changed nothing. I ran it 4 more times, and all were around 4k


to K3N:

I would overclock, but i only have the stock hsf, and if you check my thread in the cpu sub-f, you'll see that my temps are going crazy

 

[Stumps]

interesting....I went for a A64@2.87ghz to a E4300 and had nearly a 700 point increase without overclocking the E4300...with overclocking it jumped another 800 or so points.

although I should add that my A64 is only a single channel Skt 754 setup

 

[yh125d]

hmm. are 754 a64's inherently worse than 939 a64's?

how similar was the 754 rig to your e4300 rig. same ram amount?

 

[postmortemIA]

Originally posted by: yh125d
hmm. are 754 a64's inherently worse than 939 a64's?

how similar was the 754 rig to your e4300 rig. same ram amount?

in a nutshell, it is same as running 939 one in the single channel mode.

 

[Stumps]

Originally posted by: yh125d
hmm. are 754 a64's inherently worse than 939 a64's?

how similar was the 754 rig to your e4300 rig. same ram amount?

I recycled my ram and Video card from my old A64/NF3 setp

Both systems used 1GB DDR400 (2x 512mb) and a Sapphire X1950Pro 512mb AGP

 

[anandtechrocks]

Looking at that 3D Marks page you linked, it looks like you're running your RAM in single channel mode (instead of dual channel) which will hurt your score.

 

[yh125d]

Its a single 1gb stick. I was running 3 x 512 in my old system in single channel

 

[gobucks]

dude, the stock heatsink isn't THAT bad - the chip only runs at like 1.3V, when I OC'd my e6400 i didn't even need to touch the voltage until i hit 3GHz. Temps don't start getting out of control until you start bumping the vcore up past 1.45V. I'd at least try to hit 2.6GHz at stock voltage. that will give you a better comparison on the performance improvement of a C2D over an A64.

And yes, the RAM is part of your problem, for 2 reasons - a) it is running in single channel and b) the bus on the e4300 is really slow - it's really only fast enough to support DDR2-400 at a 1:1 ratio - your bus is only running at 60% of the speed of your RAM. C2Ds are much more memory bandwidth-hungry than A64s, which really only cared about low memory latency. Adding another stick of memory and/or OCing your FSB to take advantage of the faster memory will help out quite a bit. if neither is an option, try slowing down your memory and seeing if you can reduce your memory timings - running your RAM in 1:1 with low timings will produce better results than running your memory faster than the bus with high timings.

 

[yh125d]

I'm running my mem at 3:5 atm. I know how to get it to run at 1:1, but I can't find anything for the timings. Is it in the BIOS? What timings should I shoot for?

 

[Stumps]

Originally posted by: yh125d
Its a single 1gb stick. I was running 3 x 512 in my old system in single channel

that's your problem right there....C2D's perform much better in Dual channel mode...they love memory bandwidth.

 

[ribbon13]

Yeah, but he did the right thing, because being stuck with 2x512mb and wondering how a 1GB stick would interact is not fun and 4x512mb isn't great either.