G80 with A64, X2, AM2, 939 results

[CAD4466HK]

It seems that us A64 users might want to reconsider or OC the frell out of our
CPU's before we upgrade to the G80's:mad:

According to the non-DX10 benchmarks, at least a 2.4Ghz for the 8800GTS
and at least 2.6Ghz for the 8800GTX to be worth the upgrade.

It seems a C2D might be on my Christmas list this year:p

I can only imagine how bad it will be when the DX10 titles paired with Vista
come out;)

It's going to take more then just having a 750w-1kw PSU to drive these monsters to the max:mad: {meaning quad-core, 4x4}
http://www.firingsquad.com/hardware/geforce_8800_gtx_gts_amd_cpu_scaling/

[kuhla]

So I guess I am pretty much ok for now with my 2.75ghz Opteron 165 (I will be pushing for higher once I get my watercooling setup put together) :) I run 2.8ghz in the cooler months and all the fans on full (usually only at 1/2 speed for noise). I might be using my evga step-up for a GTX or GTS eventually.

[CAD4466HK]

I wish I "could" step up:mad:

[Bbq]

I should be ok with a 3 ghz 146 at 1.4 v.

[killnine]

I am fairly confident that DirectX 10 will actually aleviate the CPU bottlenecking issues. If you read up on it, the new D3d version is set to reduce ops send to graphics subsystems like the CPU in an effort to keep most of the calcs on the gpu.

However, this is only the "goal". Reality may be something entirely different.

[Manny Calavera]

I wish I "could" step up:mad:



I just sent off my GTS back to eVGA to step up to a GTX.Next time grab a eVGA,and you will be able to do a step up as well :D

[CAD4466HK]

I just sent off my GTS back to eVGA to step up to a GTX.Next time grab a eVGA,and you will be able to do a step up as well :D

I've had 5 eVGA 7900gt KO's and after I bought the first one thru Newegg,and
RMA'ed it thru eVGA the second time, I'm not able to step up anymore since
July, trust me, I would rather "down" step to a working 7800GTX:p

[Manny Calavera]

I've had 5 eVGA 7900gt KO's and after I bought the first one thru Newegg,and
RMA'ed it thru eVGA the second time, I'm not able to step up anymore since
July, trust me, I would rather "down" step to a working 7800GTX:p



Wow,sorry to hear about your troubles with your system :(

[soltys]

That really depends if the game is or isn't limited by your current cpu power. Besides, even if you don't get much higher framrates in some older title (on your old cpu), you can still run them now with all the bells and whistles of x16 af / x4+x16 aa in 1600x1200 or higher :)

[SKYMTL]

Actually, I wrote a huge thread about this over on NCIX. The difference between AM2 and the Conroes at stock speed is next to nothing when it comes to higher-resolution gaming. Do you notice that they didn't include any Conroes in those benchmarks? I wonder why. Here is a picture I made which includes a few outtakes from this AMD article and their original 8800 test with a X6800:

http://img219.imageshack.us/img219/4879/5f5a60d88b261320dd59bf7ux7.th.jpg (http://img219.imageshack.us/my.php?image=5f5a60d88b261320dd59bf7ux7.jpg)

Anyone else notice that the scores are virtually identical? This goes to show that the who myth of a AMD X2 processor bottlenecking a G80 is quite a bunch of BS. The X2 4200+ scores are virtually identical to the scores achieved with the X6800 when using the 8800GTS.

For more info about this and counter arguments:
http://forum.ncix.com/forums/index.php?mode=showthread&forum=215&threadid=1223720&product_id=0&forumid=215&pcbuilderid=0

[jonnyGURU]

I can confirm this as well.

When running 3DMark 2006, my FX62 didn't do any better or worse than either the X6800 or QX6700 builds with the same cards in them. I could switch out the video cards, and the score would change, but still no "CPU bottleneck" that I could see.

For the record; the Quad Father with a pair of 8800GTS's scored higher than the FX62 and QX6700 or X6800 with a pair of GTX's, so obviously the CPU still plays a HUGE part, but then again we're only talking about one particular benchmark that I'm using (because I'm lazy.)

[SKYMTL]

Bingo :)
And the difference should be even less when gaming.

[madmat]

With 3DMark '06 I ignore the CPU score altogether and look at the SM 2.0 and 3.0 scores instead. I've seen 7950GT scores on Conroes that are higher overall than mine but had lower SM scores. The CPU's are artificially inflating what should be a graphics intensive score.

[Boyo]

Thanks for setting the record straight. That helps clear things up.