3200+ 90NM (939) --REVIEWED AND COMPARED

[Bugler]

Nice written review posted today.

It compares the four processors mentioned above with many benchmarks.

Text

 

[BW86]

cool, nice read

 

[Painman]

Hmm, that chip is looking tastier and tastier.

 

[Caveman]

OK... I'm confused...

When the 90nm fiirst came out, it seems like AT was reporting that the chip might actually run HOTTER than its 130nm counterpart... This review seems to indicate that the chip is faster at any given clock speed than the 130 nm 3500+ Is this for real? Then... can I assume that a 90nm 3500 might be as stout as a 130nm3800+

Please help me understand 90 vs 130 nm technology...

 

[flexy]

more and more i have a feeling that AMD's PR rating just plain SUCKS.

The benchmark results are (IN MY OPINION) neglectible, since i dont care whether i get 88,7 FPS or 90 FPS in a game....and for 15 pts more (of 22000) or so you cant buy yourself an icecrean either. This is benchmarking nonsense at its best.

HOWEVER.....i think its disturbing that a CPU named "3500" is obviously a tad slower than the 'slower' rated one. Ok, at same speeds. But if BOTH chips have the alleged "same" new optimizations...well then the 3500 90nm has em as well as the 3200 90nm.....

 

[Bugler]

Originally posted by: flexy
more and more i have a feeling that AMD's PR rating just plain SUCKS.

The benchmark results are (IN MY OPINION) neglectible, since i dont care whether i get 88,7 FPS or 90 FPS in a game....and for 15 pts more (of 22000) or so you cant buy yourself an icecrean either. This is benchmarking nonsense at its best.

HOWEVER.....i think its disturbing that a CPU named "3500" is obviously a tad slower than the 'slower' rated one. Ok, at same speeds. But if BOTH chips have the alleged "same" new optimizations...well then the 3500 90nm has em as well as the 3200 90nm.....

The review indicated they were testing against the .13 version of the 3500+, not the 90nm version.

1) Athlon FX53 (S939 - 2.4ghz ? 1Mb L2 cache - 0.13µ)
2) Athlon64 3500+ (S939 ? 2.2ghz ? 512k L2 cache ? 0.13µ)
3) Athlon64 3400+ (S754 ? 2.2ghz ? 1Mb L2 cache ? 0.13µ)
4) Athlon64 3200+ (S939 ? 2ghz ? 512k L2 cache - 0.09µ)

 

[flexy]

n/m...see above...yes..his 3500 is a 130nm one...

 

[Caveman]

Anyone know the answers to my questions above?

 

[Zebo]

Originally posted by: flexy
more and more i have a feeling that AMD's PR rating just plain SUCKS.

The benchmark results are (IN MY OPINION) neglectible, since i dont care whether i get 88,7 FPS or 90 FPS in a game....and for 15 pts more (of 22000) or so you cant buy yourself an icecrean either. This is benchmarking nonsense at its best.

HOWEVER.....i think its disturbing that a CPU named "3500" is obviously a tad slower than the 'slower' rated one. Ok, at same speeds. But if BOTH chips have the alleged "same" new optimizations...well then the 3500 90nm has em as well as the 3200 90nm.....

Make you own PR rating. It works like this since the skt 754 3000 is closest to the 3.0 northwood in all the reviews I've seen.

A64 speed x 1.5 = P4C
Add 5% for Dual Channel
Add 5% for 1 mb lvl 2
Subtract 2% for skt 940
AXP-speed x 1.25 =P4C

EX?
How fast is 3800? = 2400mhz x 1.5 x 1.05 = 3780mhz P4

or 3780/1.25 = 3024Mhz AXP

How fast is a 2800 skt 754? 1800Mhz x 1.5 = 2700 mhz P4C

Which youll also find in the reviews a 2800 is a touch slower than the P4C 2.8 too.

Try these multipliers for this review or any comparison review after all benches are normalized and I think You'll find it fairly accurate.

 

[Concillian]

When overclocked the 3400+ suffers, lack of ram bandwidth as this platform is only single channel and is limited to 208mhz /DDR416 due to ram and board limitations with the ram used.

WTF? Why did they cripple the S754 processor by using RAM that wasn't up to the task? Why not use the same RAM they did on the S939 that handled 250 MHz just fine?

I don't get it.

 

[Zebo]

This guy was a philosophpy major tuned computer reviewer... geez he talked a lot. Would have been a much better review to have a 3.7E , also possible wih overclcoking, in there to compare.

 

[Ronin]

The 754 is crippled because of the single on-die memory controller. No type of RAM is going to overcome that.

 

[Bugler]

I had been leaning towards the 3500+ 130 based upon the heat concerns raised by others. However, this review has me eyeing the 3200+ 90nm.

 

[iwantanewcomputer]

interesting to see the 90 nm hanging with the 130 nm fx clock for clock, now we need a 90 m fx

 

[Kunibert]

Have you read this?
"System testing was tailored to provide comparative data in relation to stock and overclocked speeds of the 3200+ 0.09µ processor. CPU multipliers for faster running chips were lowered to 10x to bring operating clock speeds in line for each chip and provide better contrast to performance differences. Benchmarks were performed at 2 GHz and 200 MHz HTT, default clock speed for 3200+ A64 and again at 2.5 GHz and 250 MHz HTT for each of the CPUs."

He testet the chips at the same speed to check the temperature etc. and not how fast they are at Stock Speed

 

[Bugler]

Originally posted by: Kunibert
Have you read this?
"System testing was tailored to provide comparative data in relation to stock and overclocked speeds of the 3200+ 0.09µ processor. CPU multipliers for faster running chips were lowered to 10x to bring operating clock speeds in line for each chip and provide better contrast to performance differences. Benchmarks were performed at 2 GHz and 200 MHz HTT, default clock speed for 3200+ A64 and again at 2.5 GHz and 250 MHz HTT for each of the CPUs."

He testet the chips at the same speed to check the temperature etc. and not how fast they are at Stock Speed

Yes, I had read that. However, what I don't understand is even under those situations, how the 3500+ could look so bad verses the rest of the field.

 

[flexy]

Originally posted by: Bugler

Originally posted by: Kunibert
Have you read this?
"System testing was tailored to provide comparative data in relation to stock and overclocked speeds of the 3200+ 0.09µ processor. CPU multipliers for faster running chips were lowered to 10x to bring operating clock speeds in line for each chip and provide better contrast to performance differences. Benchmarks were performed at 2 GHz and 200 MHz HTT, default clock speed for 3200+ A64 and again at 2.5 GHz and 250 MHz HTT for each of the CPUs."

He testet the chips at the same speed to check the temperature etc. and not how fast they are at Stock Speed

Yes, I had read that. However, what I don't understand is even under those situations, how the 3500+ could look so bad verses the rest of the field.

the 3500 does NOT look bad. Dont forget that his is a 130nm part, the 3400 had a LARGER cache...etc..etc... his 3500/130nm test is totally un-interesting.

He should have rather tested with the new 90nm part - i wouldnt be surprised seeing the 90nm part wiping the floor with the others.

Also...i would really like to KNOW about overclockability and temps...thats important too !

 

[Glavinsolo]

3500+ 90nm

This was a thread that I started when I recieved my 3500+ and played with it for a few days.
Note: I was running SATA on the southbridge and that is what caused my errors I am in the middle of a storage rearrage to clear up SATA on the southbridge. I will post back with my drives on the locked PCI bus

-Glavin

 

[Concillian]

Originally posted by: Kunibert
Have you read this?
"System testing was tailored to provide comparative data in relation to stock and overclocked speeds of the 3200+ 0.09µ processor. CPU multipliers for faster running chips were lowered to 10x to bring operating clock speeds in line for each chip and provide better contrast to performance differences. Benchmarks were performed at 2 GHz and 200 MHz HTT, default clock speed for 3200+ A64 and again at 2.5 GHz and 250 MHz HTT for each of the CPUs."

He testet the chips at the same speed to check the temperature etc. and not how fast they are at Stock Speed

But then he went and gave the S754 processor inferior memory that only ran 208MHz while all the S939 processors ran 250MHz. On top of that, the S754 processor was tested in DirectX 9.0B instead of 9.0C.

The 939 comparisons appear to be valid, but the S754 vs. 939 comparison is questionable. I would have just left it out of the review with so many differences.

 

[flexy]

>>>
2.75 on air hmm, and 3.0 on water it looks like.
>>>

oh.....ooohhhhhhh.....ohhhhhhhhh......... (resist pressing the 'order button" on newegg.... )
I heard prices will drop mid october....have to wait a bit 'til then and nforce 4.....whow.....2.7 on air ...


btw. i plan to get the Zalman AlCu...

 

[flexy]

He testet the chips at the same speed to check the temperature etc. and not how fast they are at Stock Speed

and to validate/invalidate the claim that 0.90 parts have optimizations over the 130nm parts.

 

[SonicIce]

I need a reliable site that benches the 3200+ 754 with the 3200+ 939 at stock speeds.