barton 2500 vs 1700+ at 200fsb performance diff

[us1]

getting ready to buy, after reading the forums and some impressive overclocks, would you tell me how much performance difference there would be between an barton 2500 + running at 200x11.5 and an 1700+ running at 200x11.5. I know both will overclock more but all things being equal except the extra cache, how much differance is there performance wise. my first post here so cut me some slack, thanks

 

[Corporate Thug]

on average, the barton performes 5% better

 

[Jeff7181]

Just look at the benchmarks for the XP2800 Barton... the XP2600 runs at the same speed.

 

[4x4expy]

Personally I would say that if you already have a Tbred doing 200x11.5 it is probably not worth the $90 to change. If you don't have either, spend the extra $40 and get the Barton.

For a comparison here is my PCMark2002 scores on cpu and mem with a 1700+ at 12x200. You can loosely compare to Jeff's scores above(Albeit diff clock speeds) But the mem score is at the Same FSB, making it an apples to apples comparison, so you can see the diff in the 2.

CPU: 7244 MEM: 5683

(Maybe in a little while, change the multi to 11 and give an updated score, to give a good CPU score comparo)

 

[Jeff7181]

For some more comparison... I had benchmarked my CPU at 2.3 Ghz and got about 7100 CPU marks. But it's not stable at that speed.

 

[4x4expy]

Update: Ok, I rebooted at 11x200 and re benched with pcm2002.


CPU: 6774 MEM:5590

Now that can be compared to the Barton at the same speed. The CPU scores are a dead heat, and there is a definite memory advantage due to the increased cache, but this doesn't really tell you how much perceived performance increase you will get. This is only meant as a rough basic comparo.

 

[us1]

Thanks for the great answers and information on the comparison of the 2 cpu's. So, do you think it is worth the extra 40 bucks approx. to get the barton?

 

[MDE]

On my 2500 @ 200x11 I got 6719 CPU and 5754 memory

 

[MaDDaWg1018]

hmmm...well, i currently have an athlon tbird 266 fsb @ 1.33 stock, and i just ordered a barton 2500+. i was also debating on whether i should shell out a little extra for it, and well, the 512k l2 cache just sold me. a lil extra cache never hurts. =) but as far as "feeling" the difference...i dont think it'll be that noticeable, if at all.

 

[4x4expy]

Originally posted by: MonkeyDriveExpress
On my 2500 @ 200x11 I got 6719 CPU and 5754 memory

Interesting... You're mem score is much closer to mine than Jeff's... Maybe the extra cache isn't completely responsible for the 500-600 point diff. It seems some of the difference is coming from sc/dc and ram timings. I am running sc ddr400 2.5,3,3,6. Jeff is obviously at DC, probably tighter timings since I think the Palmtree has BH-5 chips. This makes me feel that maybe there is even less difference than I previously thought. I bought a pair of Corsair 256mb pc3200ll platinum sticks off of the FS forum, and they will hopefully be here today. I will repost my scores with the DC mem at the tightest timings I can get stable to see how much it pulls my score up.

Originally posted by: us1
Thanks for the great answers and information on the comparison of the 2 cpu's. So, do you think it is worth the extra 40 bucks approx. to get the barton?

Honestly you won't regret whichever you buy. The Barton does have a performance edge over the Tbred. It may be noticable in some apps, and not in others. But if in any app there is an advantage it will always favor the Barton. I like my Tbred because I just think it is silly to get such performance out of a $50 chip! I also have another rig with a p4 2.4b that runs 2.95ghz on stock voltage, and with all same components my cheap Tbred (@ 2.4ghz) is faster in 95% of apps!

 

[4x4expy]

Corsair 256mbx2 installed! SPD timing of 2-2-2-6 at ddr400!(I haven't tightened beyond that yet)

1700+ @ 200x12

CPU: 7357 MEM: 6047

So the DC didn't make up all of the diff of course. Off now to run some other benchies to see what kind of performance gain I got.


Edit:
WOW!!! My 3dmark01 score teeters around 16900 and has peeked above(slightly) 17k, with tweaks a couple times. With this ram configuation, I ran the loop 1 time and BAM!! 17866 marks! I don't think the 9700p is maxxed either, so I may be able to tweak it above 18k. Not to mention this mem may let me be stable at higher fsb. Back to the drawing board!

 

[Jeff7181]

Those of you with RAM scores down in the middle 5000's... are you running dual channel?

 

[bjc112]

Originally posted by: Jeff7181
For some more comparison... I had benchmarked my CPU at 2.3 Ghz and got about 7100 CPU marks. But it's not stable at that speed.

I scored 7000+ ( somewhere in that 7156 or something ) @2.4ghz for comparison...

EDIT: 7226

 

[bjc112]

Originally posted by: Jeff7181
Those of you with RAM scores down in the middle 5000's... are you running dual channel?

I am.

Does seem a bit low... Sandra scores seem on par with 3000/3000

 

[ForsakenMaster]

Yes it is

 

[4x4expy]

Originally posted by: bjc112

Originally posted by: Jeff7181
Those of you with RAM scores down in the middle 5000's... are you running dual channel?

I am.

Does seem a bit low... Sandra scores seem on par with 3000/3000

Those sandra scores are about right for ddr400.

 

[MDE]

I'm running single channel with pretty loose timings on cheapo memory.

 

[Duvie]

Here is the skinny...At those cpu settings listed the only difference is the Barton has the 512kb of L2 cache....However that may make 5% in some apps but there are some apps that make no difference at all.... I would stay at what you are at....