X6800 review out early, with overclocking results!

[Your_Boss]

I couldn't find a thread on this site yet, so here it is, wow!






Just look at the FEAR benchmarks!

Here is the entire link, it's in French, but I Googled it for English. It's a good read.
http://translate.google.com/translat...language_tools

Some pricing too.

Xeon 5160 3.00 GHz, FSB1333, 4 Mo L2, socket LGA771 - 851 US dollars
Xeon 5150 2.66 GHz, FSB1333, 4 Mo L2, socket LGA771 - 690 US dollars
Xeon 5140 2.33 GHz, FSB1333, 4 Mo L2, socket LGA771 - 455 US dollars
Xeon 5130 2.00 GHz, FSB1333, 4 Mo L2, socket LGA771 - 316 US dollars
Xeon 5120 1.86 GHz, FSB1066, 4 Mo L2, socket LGA771 - 256 US dollars
Xeon 5110 1.60 GHz, FSB1066, 4 Mo L2, socket LGA771 - 209 US dollars
Core 2 Duet X6800 2.93 GHz, FSB1066, 4 Mo L2, socket LGA775 - 999 US dollars
Core 2 Duet E6700 2.66 GHz, FSB1066, 4 Mo L2, socket LGA775 - 530 US dollars
Core 2 Duet E6600 2.40 GHz, FSB1066, 4 Mo L2, socket LGA775 - 316 US dollars
Core 2 Duet E6400 2.13 GHz, FSB1066, 2 Mo L2, socket LGA775 - 241 US dollars
Core 2 Duet E6300 1.86 GHz, FSB1066, 2 Mo L2, socket LGA775 - 209 US dollars
Core 2 Duet E4200 1.60 GHz, FSB800, 2 Mo L2, socket LGA775 - Price nonavailable
Core 2 Duet T7600 2.33 GHz, FSB667, 4 Mo L2, socket mPGA479 - 637 US dollars
Core 2 Duet T7400 2.16 GHz, FSB667, 4 Mo L2, socket mPGA479 - 423 US dollars
Core 2 Duet T7200 2.00 GHz, FSB667, 4 Mo L2, socket mPGA479 - 294 US dollars
Core 2 Duet T5600 1.83 GHz, FSB667, 2 Mo L2, socket mPGA479 - 241 US dollars
Core 2 Duet T5500 1.66 GHz, FSB667, 2 Mo L2, socket mPGA479 - 209 US dollars

[hipogtr]

Core Duo - simply too powerful.

[Nathanael]

Amazing stuff =D .

[doc6886]

Dude... the X6800 at stock is unbelievable. Imagine 5ghz like people have already achieved. I'd like to see the game benchies with 5ghz...

[DaWaN]

Core 2 Duet

[informal]

You couldn't fine a topic about this review?Man there were at least 2 of them...
Search harder next time.And the test is not showing us anything we didn't know earlier...

[[ypher]

I'd like to see at least 10x7 benches with AA/AF, 640x480 seems worthless.

[vengance_01]

I'd like to see at least 10x7 benches with AA/AF, 640x480 seems worthless.

not really, at that res we see big differences as the cpu's do alot of the work. its a decent way to show off the power of a cpu.

[J-Mag]

not really, at that res we see big differences as the cpu's do alot of the work. its a decent way to show off the power of a cpu.

Running test at super low resolutions is just like running a synthetic bench. It shows off, but doesn't mean anything for real world performance.

For games CPUs just don't matter that much these days...

[Turtle 1]

Running test at super low resolutions is just like running a synthetic bench. It shows off, but doesn't mean anything for real world performance.

For games CPUs just don't matter that much these days...

Strange you feel that way I have a differant take on what I am seeing inthat review.

If you would look at the O/C section for gaming. You will notice that Conroes keep scaling up. Even at 3.6 ghz. While the AMD @ 3.1 loses groung and infact runs slower than a 2.7 amd.

I believe that when the R600 are released and If you run XF the Conroe will show off a lot more power in gaming than AMD . AT high res and min frames the differances should be rather LARGE.

[Celeron Gamer]

I'd like to see at least 10x7 benches with AA/AF, 640x480 seems worthless.

The reason they chose 640X480 is because they do not pose enough stress for a bottleneck of GPU, since Far Cry less demanding than others they can reach up to 800X600 without problem.

If you were to run it at high res, they all jam in from the bottleneck, you'd see the Core 2 Duo X6800 to be only 2fps higher than a Opteron 165, or so on

[[ypher]

I play games a lot, 1280x1024 is my favourite res. and these kind of tests I can name "real life". I'm waitin' for them

[Celeron Gamer]

I play games a lot, 1280x1024 is my favourite res. and these kind of tests I can name "real life". I'm waitin' for them

But if your GPU struggles at 1280X0124, you won't notice the difference between Conroe, and a AMD Opteron.

As I've said before, 10*7 would be pointless in terms of CPU benches, as it would not show pure processing power, but biased against the best performing CPU by hitting a fps roof from the GPU.

[[ypher]

I know, I know. I'm just curious how it performs

[kiwi]

How can dothan @ 2.8Ghz beat Yonah at 3Ghz in Fear? RAM taimings must be way off

But sure Conroe looks great

[Torin]

But if your GPU struggles at 1280X0124, you won't notice the difference between Conroe, and a AMD Opteron.

As I've said before, 10*7 would be pointless in terms of CPU benches, as it would not show pure processing power, but biased against the best performing CPU by hitting a fps roof from the GPU.

Depends on what AA/AF settings his GPU struggles at. Even at lower resolutions, the CPU is a bottleneck on all the most recent games, so all will benefit from a Conroe, some more than others.

[Vric]

agree.. they are testing a CPU, not a video card. Maxing all the eye candy and high resolution will saturate the video card and the CPU will only work at a fraction of his potential. If you want to see bench with 1600x1200 resolution, you will see that in a Video card review, not a CPU review.

[[ypher]

With QuadSLI I wouldn't be so sure

[Jumba]

[ypher, take a look at this thread for some hi-res game benchmarks.

[[ypher]

Many thanks, Jumba. Looks like a week of reading

[Okda]

how the 165 @ 2.7 will score more than the FX-62 @ 2.8Ghz

really i don't trust such review unless someone can give me a resonable explanation

for sure Conroe is very nice and thiis is not what i am talkigng about , i am only worried about the ability of this review to show true results from the AMD results

[snowwie]

Some pricing too.

Xeon 5160 3.00 GHz, FSB1333, 4 Mo L2, socket LGA771 - 851 US dollars
Xeon 5150 2.66 GHz, FSB1333, 4 Mo L2, socket LGA771 - 690 US dollars
Xeon 5140 2.33 GHz, FSB1333, 4 Mo L2, socket LGA771 - 455 US dollars
Xeon 5130 2.00 GHz, FSB1333, 4 Mo L2, socket LGA771 - 316 US dollars
Xeon 5120 1.86 GHz, FSB1066, 4 Mo L2, socket LGA771 - 256 US dollars
Xeon 5110 1.60 GHz, FSB1066, 4 Mo L2, socket LGA771 - 209 US dollars
Core 2 Duet X6800 2.93 GHz, FSB1066, 4 Mo L2, socket LGA775 - 999 US dollars
Core 2 Duet E6700 2.66 GHz, FSB1066, 4 Mo L2, socket LGA775 - 530 US dollars
Core 2 Duet E6600 2.40 GHz, FSB1066, 4 Mo L2, socket LGA775 - 316 US dollars
Core 2 Duet E6400 2.13 GHz, FSB1066, 2 Mo L2, socket LGA775 - 241 US dollars
Core 2 Duet E6300 1.86 GHz, FSB1066, 2 Mo L2, socket LGA775 - 209 US dollars
Core 2 Duet E4200 1.60 GHz, FSB800, 2 Mo L2, socket LGA775 - Price nonavailable
Core 2 Duet T7600 2.33 GHz, FSB667, 4 Mo L2, socket mPGA479 - 637 US dollars
Core 2 Duet T7400 2.16 GHz, FSB667, 4 Mo L2, socket mPGA479 - 423 US dollars
Core 2 Duet T7200 2.00 GHz, FSB667, 4 Mo L2, socket mPGA479 - 294 US dollars
Core 2 Duet T5600 1.83 GHz, FSB667, 2 Mo L2, socket mPGA479 - 241 US dollars
Core 2 Duet T5500 1.66 GHz, FSB667, 2 Mo L2, socket mPGA479 - 209 US dollars

looks like I found my next cpu

[Okda]

getting a FSB 800 is not an idea i like specially when u don't know its price

i suggest saving more and buying the E6300 which is not that expensive AT ALL

[J-Mag]

[ypher, take a look at this thread for some hi-res game benchmarks.

That's exactly what I wanted to see. However, something concerns me.

We all know that benching at lower resolutions shows off CPU performance better, right?

Here are the results copied directly from that thread for the FEAR test:

1024x768, 4xAA 16xAF, all max:
Opteron 165 @ 3GHz: Min 60 | Avg 142 | Max 370
Conroe @ 3GHz: Min 78 | Avg 162 | Max 485

1600x1200, 4xAA 16xAF, all max:
Opteron 165 @ 3GHz: Min 25 | Avg 43 | Max 172
Conroe @ 3GHz: Min 53 | Avg 86 | Max 310

Can someone tell me why the 1024 resolution has an average difference of 12% while the 1600x1200 shows a difference of 50%? Thats a little counter intuiative don't ya think?

[Nathanael]

But if your GPU struggles at 1280X0124, you won't notice the difference between Conroe, and a AMD Opteron.

As I've said before, 10*7 would be pointless in terms of CPU benches, as it would not show pure processing power, but biased against the best performing CPU by hitting a fps roof from the GPU.

Well really having more power is always good. I don't see why you need to have such a negative stance against progress.