SOLVED Broadwell compatible with Z97 Boards ??

here's how i see the pcie 2.0,3.0, 4.0, I run gtx760's in sli, I will never use all the banwidth of any of them, if i ran quad titans, then perhaps.

caddi daddi said:

here's how i see the pcie 2.0,3.0, 4.0, I run gtx760's in sli, I will never use all the banwidth of any of them, if i ran quad titans, then perhaps.

Click to expand...

You have to educate me on this lol , maybe point me towards an article on how does this actually works , being why is Titan special and how 760 aren't ?

Well yeah I know Titan runs faster , but the Bandwith has something to do with it just because its a faster card ?

Faster card = more data is going through pcie bus. In this case there is nearly no difference if you are using PCIE 2.0 or 3.0 as even 2.0 is fast enough for high end graphics cards. In some cases you may see little drops in performance on the fastest cards but maybe 1-5% max. Nothing that you will notice.
It's also the main reason why PCIE 3.0 x8 is good enough for SLI/CF as it's as fast as PCIE 2.0 x16.

Woomack said:

Faster card = more data is going through pcie bus. In this case there is nearly no difference if you are using PCIE 2.0 or 3.0 as even 2.0 is fast enough for high end graphics cards. In some cases you may see little drops in performance on the fastest cards but maybe 1-5% max. Nothing that you will notice.
It's also the main reason why PCIE 3.0 x8 is good enough for SLI/CF as it's as fast as PCIE 2.0 x16.

Click to expand...

Ah I see , so basically if they were to release 4.0 it's more for multi gpu setup then , so i guess
FOR EXAMPLE having a " X8 X8 " on a 4.0 setup maybe yet it's still fast as a " X16 X8 " setup on the 3.0 , this makes thing very clear now , Tks for the info

But I guess for my case , I shouldn't really bother. This is more like more for super high end cards or Xfire / SLI setup. Since I don't even bother going multi GPU.

I guess even like you said even the highest end card now maybe like 290X , with 3.0 it's plenty lol.

What about Dual GPU on a Single PCB , how does that work ??
Since it be slotted on the X16 slot , will it split it into maybe like X16 X8 configuration instead or something ?? Or it will run both the GPU chip as a X16 configure ??

X16. On a not terribly related note, dual gpu has a plz chip on it for intragpu communication.

If they release pcie 4.0 then it won't help at all unless pcie lines will be split and you will be able to run 4 pcie x8 on a standard board without additional pcie lines from 3rd party chip. If they make something like 2 pcie x8 + 1 x4 then it won't give any performance improvement over pcie 3.0 as new cards are not even using full bandwidth of pcie 3.0 not to mention 4.0.

EarthDog said:

X16. On a not terribly related note, dual gpu has a plz chip on it for intragpu communication.

Click to expand...

" dual gpu has a plz chip on it for intragpu communication "



You have to dumb this down to my level , I couldn't understand at all at this sentence....could either my English or my lack of Knowledge in IT

It has a chip that makes hardware think that these 2 GPUs are working in SLI/CF and uses full bus bandwidth. So you see 1 card but system sees 2. That chip adds communication between these 2 GPUs too.
That's why boards have often 2 pcie slots but are certified for quad gpu CF/SLI.

PLX chip.. sorry. Was mobile when I wrote that and auto correct got me.

Thanks woomack.

Woomack said:

It has a chip that makes hardware think that these 2 GPUs are working in SLI/CF and uses full bus bandwidth. So you see 1 card but system sees 2. That chip adds communication between these 2 GPUs too.
That's why boards have often 2 pcie slots but are certified for quad gpu CF/SLI.

Click to expand...

Ah I see so , I guess a Dual GPU on a Single PCB has it's own advantages too.
But strangely how come individual XFire / SLI cards runs slightly faster than the ones that combined ??

But in my case I will actually prefer 2 Chip in one PCB instead of individual setup.
Saves alot more space and with proper cooling it's cooler too it seems.

I wonder how does the VRM hold out though....

They run slower because generally it is a cut down version of the two cards. Meaning either the clock speed or shaders/stream processors are not the same as two single cards. Except the 295x2 (read my front page review).

VRM's hold out just fine as there are two, one for each GPU.

EarthDog said:

They run slower because generally it is a cut down version of the two cards. Meaning either the clock speed or shaders/stream processors are not the same as two single cards. Except the 295x2 (read my front page review).

VRM's hold out just fine as there are two, one for each GPU.

Click to expand...

Good review Now if only someone could persuade Sapphire to release a 295 X2 Atomic That definitely will blow things up lol