680i or P35?

retardude48 said:
I can sufficiently say that I have tried updating the bios. But some bios work better than others. So, depending on ur system, an older bios may work better at overclocking than newer bios. But currently, I (the friend with the 650i), have the latest bios. Plus he is talking about stock voltage. You have prob. messed with voltages, and thus bought an aftermarket cooler, and ur sig. confirms it. You can def. get the same output ur getting right now on an 680i or even an P35 board. But if you are looking to overclock ur processor more than 3.0GHz, good luck with that since you're pretty much at the wall with that processor and pushing it even more will degrade the processor. If you want faster speeds, get a better processor. I'm not tryin flame you or anything but i dont see the need for a 680i when you already have a good enough Asus P35 chipset motherboard that will easily spank the 680i chipset. Also to mention, that Anandtech article was written November 8th, 2006 and the P35 chipset was released in June of 07'. But, as an honest answer to your first post, I would spend a little extra and get a P35.
Click to expand...



ok, AGAIN, SOMEONE GAVE ME A 680i! AT THEIR EXPENSE!

and my P5K-E has DIED ON ME TWICE, and I'M NOT USING THAT PIECE OF CRAP AGAIN.

and i want to try something different.

yes i messed ith voltages... to get to 3.2.

my new board was so crappy that it wouldnt even apply any voltage. so i was stuck at the stock 1.325 Vcore. that was able to run to 3.0 ghz on my E2180.

and that is NOT the limit. this processor an do 3.33 at 1.48 ish Vcore.
under water could go to 3.5 easily.

my "aftermarket cooler" was purchased when I bought all my other parts. I NEVER used the stock cooling its way too insufficient.
 
AmongTheChosenX said:
ok, AGAIN, SOMEONE GAVE ME A 680i! AT THEIR EXPENSE!

and my P5K-E has DIED ON ME TWICE, and I'M NOT USING THAT PIECE OF CRAP AGAIN.

and i want to try something different.

yes i messed ith voltages... to get to 3.2.

my new board was so crappy that it wouldnt even apply any voltage. so i was stuck at the stock 1.325 Vcore. that was able to run to 3.0 ghz on my E2180.

and that is NOT the limit. this processor an do 3.33 at 1.48 ish Vcore.
under water could go to 3.5 easily.

my "aftermarket cooler" was purchased when I bought all my other parts. I NEVER used the stock cooling its way too insufficient.
Click to expand...

Ok I have the E2160 which is the exact same as your proc with a 9x multiplier. I am currently running a 344 fsb on only 1.4vcore with my P45 board. Seems odd that you should have to go to 1.48vcore just to get 333fsb.

Also stock hsf is great. I'm using it on my proc right now and it is far from "insufficient". I idle mid 50's under load.
 
yah ok. sufficient... thats funny.
obviously youre stupid, because FSB alone has NOTHING to do with the Vcore of the processor. Vcore only applies when you put the FSB together with the multiplier for your final frequency.

first of all, I used to run a 395 FSB with an 8x Multiplier.

which gave me 3160 mhz.
and it could run this at 1.425 Vcore

i also ran 333 FSB x 10x Multiplier for 3.33 ghz. that took 1.5 Vcore
this was for benching mostly though.

secondly, I idle at 25-28 Degrees. under load of 3.2 ghz in Prime 95, i get 55 with 1.425 Vcore.

so honestly, the stock cooler sucks.

i know what i'm talking about dude/
 
AmongTheChosenX said:
yah ok. sufficient... thats funny.
obviously youre stupid, because FSB alone has NOTHING to do with the Vcore of the processor. Vcore only applies when you put the FSB together with the multiplier for your final frequency.

first of all, I used to run a 395 FSB with an 8x Multiplier.

which gave me 3160 mhz.
and it could run this at 1.425 Vcore

i also ran 333 FSB x 10x Multiplier for 3.33 ghz. that took 1.5 Vcore
this was for benching mostly though.

secondly, I idle at 25-28 Degrees. under load of 3.2 ghz in Prime 95, i get 55 with 1.425 Vcore.

so honestly, the stock cooler sucks.

i know what i'm talking about dude/
Click to expand...

lol I idle at that same temp. and you get the same load temp with prime 95 that I get at a slightly lower vcore. Still doesn't seem like the stock cooler sucks. Especially considering I have an ambient temp of 27C and only 2 case fans, and my case is stuffed under a desk with no place to exhaust but into itself, with a radeon flaming hot 4850 under it.

Secondly, I was never attempting to "take shots" at you, just explaining my experiences. You turned this into a flame war with idiodic comments like "Obviously your stupid".
 
AmongTheChosenX said:
yah ok. sufficient... thats funny.
obviously youre stupid, because FSB alone has NOTHING to do with the Vcore of the processor. Vcore only applies when you put the FSB together with the multiplier for your final frequency.

first of all, I used to run a 395 FSB with an 8x Multiplier.

which gave me 3160 mhz.
and it could run this at 1.425 Vcore

i also ran 333 FSB x 10x Multiplier for 3.33 ghz. that took 1.5 Vcore
this was for benching mostly though.

secondly, I idle at 25-28 Degrees. under load of 3.2 ghz in Prime 95, i get 55 with 1.425 Vcore.

so honestly, the stock cooler sucks.

i know what i'm talking about dude/
Click to expand...


Just get the 680i. There reliable now and you can probably get your CPU up there
 
Dan_D said:
I'm going to disagree with you there. I've got a stack of dead 680i SLI boards that says otherwise.
Click to expand...

My god man! What were you doing to those boards?:eek:

I think they are reliable now a days and some still have problems, but your pretty much safe now. Some boards can OC some cant.
 
Many people had issues with 680i mobos mainly due to the boards inability to OC ram to high frequencies because this usually meant applying voltages beyond the board's recommended setting. Anything over 2.2v applied to memory is generally accepted as "do not go there" area by those who have thoroughly tested these boards (evga forums). I've also had issues with the board when I tried to OC the ram beyond 1000mhz (I have couple of pairs of micron based ram that I know can handle crazy overclocks) or applying voltages beyond 2.2v. Errors galore in orthos and during memtest, but passes with no errors on a different board.

As long as you stay below 2.2v, you shouldn't have too much issue with this board. Again, the NB on this board does not like high ram OC. Keep the speed to reasonable OC or at stock then tighten the timings a bit for best performance.

Later versions of A1's also had their mosfets increased to 24 from the original 18 so this should help out with the board's OC potential. My board unfortunately has only 18.

Btw, I had no problems pushing my e8400 to 4ghz on this board or my e6600 to 3.6ghz, applying 1.37v and 1.5v respectively and both were fully stable at this speed.
 
3far4shot said:
My god man! What were you doing to those boards?:eek:

I think they are reliable now a days and some still have problems, but your pretty much safe now. Some boards can OC some cant.
Click to expand...

To be fair, I overclocked the crap out of them using dual and quad core processors. The systems were perfectly solid and stable from anywhere between 3 days to 3 months. After that the boards all died. Many of the dead 680i SLI boards I've had were DOA. Almost half of them were. The other half all died on me. One of them was killed by a watercooling leak, but other than that one board the rest have no excuses.

The 680i SLI still doesn't overclock quad core processors as well as other chipsets do. The 680i SLI still has voltage regulation issues and BIOS issues causing any number of issues. The most noteworthy problem is that the 680i SLI reference boards rarely work correctly with storage controllers thanks to spread spectrum settings changes in the BIOS. Let's also not get started on memory compatibility problems and lack of 45nm quad core support.

The 780i SLI is the same MCP and SPP that the 680i SLI boards used. The only difference is an added third chip that adds PCI-Express 2.0 support and 8 additional PCI-Express lanes. While these boards SHOULD have been better, they weren't. Almost none of the 680i SLI's problems were fixed on them. The 790i SLI Ultra is better, but it has data corruption issues. Though to be fair the 680i SLI and 780i SLI problems I have personally experienced and verified, while the 790i Ultra SLI problems I have not seen for myself. My experiences with the 790i Ultra SLI were actually pretty good.

Today with ATI's cards being the better deal in terms of price vs. performance and with the 4870 X2's release drawing near there is NO compelling reason in my mind to go with an NVIDIA chipset based board.
 
hdnut said:
Many people had issues with 680i mobos mainly due to the boards inability to OC ram to high frequencies because this usually meant applying voltages beyond the board's recommended setting. Anything over 2.2v applied to memory is generally accepted as "do not go there" area by those who have thoroughly tested these boards (evga forums). I've also had issues with the board when I tried to OC the ram beyond 1000mhz (I have couple of pairs of micron based ram that I know can handle crazy overclocks) or applying voltages beyond 2.2v. Errors galore in orthos and during memtest, but passes with no errors on a different board.
Click to expand...

This is not entirely true. The boards (if left on auto) would often set EPP memory modules to 2.3v on their own. Also if left on automatic actual voltage readings would often spike past 2.3v if you overclocked the board enough. This is what lead to most of the fried memory modules out there. 680i SLI reference boards all handled EPP capable memory rated for over 1000MHz speeds well in my experience. At least this was the case with OCZ and Corsair memory modules I used. ASUS, Gigabyte, etc... did not. Also all of the 680i SLI boards didn't handle 4 DIMM configurations worth a shit. In some cases you had to leave the memory at DDR2 800MHz or DDR2 667MHz speeds in order to have any kind of stability with them.

hdnut said:
As long as you stay below 2.2v, you shouldn't have too much issue with this board. Again, the NB on this board does not like high ram OC. Keep the speed to reasonable OC or at stock then tighten the timings a bit for best performance.
Click to expand...

Depends on the RAM. Micron D9 memory worked fine for me for the most part at 2.3v, but most of the time 2.1v was sufficient. Though as you approached 1200MHz the extra voltage was needed more often than not. It depended on a combination of the BIOS version and the individual board in my experience.

hdnut said:
Later versions of A1's also had their mosfets increased to 24 from the original 18 so this should help out with the board's OC potential. My board unfortunately has only 18.
Click to expand...

I have never seen this. All the boards I've ever encountered only had 18. All of the A1's I've had including a EVGA Black Pearl croaked on me in the three month time frame. (Though I'll grab my 680i SLI's and see if any have 24, but I doubt they do.)

hdnut said:
Btw, I had no problems pushing my e8400 to 4ghz on this board or my e6600 to 3.6ghz, applying 1.37v and 1.5v respectively and both were fully stable at this speed.
Click to expand...

Try some quad core overclocking. Nothing seems to kill 680i SLI motherboards as quickly as a quad core processor. Though I have seen dual core equipped 680i SLI boards die in the same time frame as those running quad cores.
 
Dan_D said:
This is not entirely true. The boards (if left on auto) would often set EPP memory modules to 2.3v on their own. Also if left on automatic actual voltage readings would often spike past 2.3v if you overclocked the board enough. This is what lead to most of the fried memory modules out there. 680i SLI reference boards all handled EPP capable memory rated for over 1000MHz speeds well in my experience. At least this was the case with OCZ and Corsair memory modules I used. ASUS, Gigabyte, etc... did not. Also all of the 680i SLI boards didn't handle 4 DIMM configurations worth a shit. In some cases you had to leave the memory at DDR2 800MHz or DDR2 667MHz speeds in order to have any kind of stability with them.
Click to expand...

With my board, the EPP setting on my OCZ 1066 would never show up as 1066 but as 667 and at 1.8v. Turns out that I had to manually configure the bios to make sure it was set at "SLI memory enabled". But, this didn't resolve the issue because either the bios or the faulty EPP setting would not give enough voltage to the memory modules and it would reset to default JEDEC settings at 667. I resolved this issue by using non-EPP memory and just OC'd the memory on my own. EPP is useless, imo, as any decent OC'ing RAM can be manually set to desired speed and voltages. For instance, the EPP setting for OCZ 1066 would automatically set the voltages at 2.3v when they ran perfectly fine at 2.1v. I'm surprised that you ran your memory at "auto" voltages. I never did this and always manually configured mine since I knew that these boards disliked anything above 2.2v. Also, I had never had issue running 4 DIMMs on my board (maybe I'm lucky). What I do know is that you need to raise the SSP voltages to about 1.4 - 1.45v to be stable with 4 DIMMs.

Depends on the RAM. Micron D9 memory worked fine for me for the most part at 2.3v, but most of the time 2.1v was sufficient. Though as you approached 1200MHz the extra voltage was needed more often than not.
Click to expand...

I mean this is a given. Higher frequencies generally requires more voltage. The problem was these OC'd memory modules requiring upwards of 2.2v which would render these boards unstable. So achieving high RAM OC's is not really feasible with these boards. Many have ran their memory fine around 1000mhz (some to even 1100's) or so but generally anything higher than that is not advised with these boards due to board's inability to be stable, possibly frying your mobo, above 2.2v. This is the main reason why I said to stay below 1000mhz. Just an advice and of course YMMV.

I have never seen this. All the boards I've ever encountered only had 18. All of the A1's I've had including a EVGA Black Pearl croaked on me in the three month time frame.
Click to expand...

http://www.evga.com/forums/tm.asp?m=78381

If you read the post, it tells you how many mosfets this particular 680i has. You can also count them yourself in the pics provided.

Try some quad core overclocking. Nothing seems to kill 680i SLI motherboards as quickly as a quad core processor. Though I have seen dual core equipped 680i SLI boards die in the same time frame as those running quad cores.
Click to expand...

http://www.evga.com/forums/tm.asp?m=123659&mpage=1&key=mosfet

Many have ran 65nm with success and with decent overclocks on these 680i's. The main problem are the 45nm quads which can't be run worth a monkey's doo-doo.

Since owning my 680i since March of 2007, I had never had to RMA it once. I have brought it back from what I thought at first was death or fried memory modules (crap load of "C1" errors or "--" or "D1" and bunch of others). Neither was true. The board is, however, very picky with the type of memory you use. To OP, I highly recommend sticking with ddr2 800, preferably micron d9's, and OC'ing these memory to much higher frequencies. Reason I state this is because the only time I had issues with memory was due to garbage EPP settings that the board did not like, which would either give off the "C1" code and refusing to boot or automatically set the voltages way too high.
 
Dan_D said:
To be fair, I overclocked the crap out of them using dual and quad core processors. The systems were perfectly solid and stable from anywhere between 3 days to 3 months. After that the boards all died. Many of the dead 680i SLI boards I've had were DOA. Almost half of them were. The other half all died on me. One of them was killed by a watercooling leak, but other than that one board the rest have no excuses.

The 680i SLI still doesn't overclock quad core processors as well as other chipsets do. The 680i SLI still has voltage regulation issues and BIOS issues causing any number of issues. The most noteworthy problem is that the 680i SLI reference boards rarely work correctly with storage controllers thanks to spread spectrum settings changes in the BIOS. Let's also not get started on memory compatibility problems and lack of 45nm quad core support.

The 780i SLI is the same MCP and SPP that the 680i SLI boards used. The only difference is an added third chip that adds PCI-Express 2.0 support and 8 additional PCI-Express lanes. While these boards SHOULD have been better, they weren't. Almost none of the 680i SLI's problems were fixed on them. The 790i SLI Ultra is better, but it has data corruption issues. Though to be fair the 680i SLI and 780i SLI problems I have personally experienced and verified, while the 790i Ultra SLI problems I have not seen for myself. My experiences with the 790i Ultra SLI were actually pretty good.

Today with ATI's cards being the better deal in terms of price vs. performance and with the 4870 X2's release drawing near there is NO compelling reason in my mind to go with an NVIDIA chipset based board.
Click to expand...


What if i put a 4870 X2 on my Nvidia chipset?

I agree about the Quad core OCing on these boards.
 
hdnut said:
With my board, the EPP setting on my OCZ 1066 would never show up as 1066 but as 667 and at 1.8v. Turns out that I had to manually configure the bios to make sure it was set at "SLI memory enabled". But, this didn't resolve the issue because either the bios or the faulty EPP setting would not give enough voltage to the memory modules and it would reset to default JEDEC settings at 667. I resolved this issue by using non-EPP memory and just OC'd the memory on my own. EPP is useless, imo, as any decent OC'ing RAM can be manually set to desired speed and voltages. For instance, the EPP setting for OCZ 1066 would automatically set the voltages at 2.3v when they ran perfectly fine at 2.1v. I'm surprised that you ran your memory at "auto" voltages. I never did this and always manually configured mine since I knew that these boards disliked anything above 2.2v. Also, I had never had issue running 4 DIMMs on my board (maybe I'm lucky). What I do know is that you need to raise the SSP voltages to about 1.4 - 1.45v to be stable with 4 DIMMs.
Click to expand...

North bridge voltage adjustments never helped me get 4 DIMMs running stable. Even at 800MHz without overclocking it was a no go on all my 680i SLI boards. EPP settings also worked perfectly on all my boards. (Aside from the voltages sometimes needing to be set to manual.) Using my OCZ Flex XLC PC2-9200 DDR2 1150MHz, the EVGA 680i SLI boards always set the modules to 1143MHz using the correct timings and voltages. Though there was some variance between boards. Some of them set the vDIMM to 2.1v or 2.3v.

hdnut said:
I mean this is a given. Higher frequencies generally requires more voltage. The problem was these OC'd memory modules requiring upwards of 2.2v which would render these boards unstable. So achieving high RAM OC's is not really feasible with these boards. Many have ran their memory fine around 1000mhz (some to even 1100's) or so but generally anything higher than that is not advised with these boards due to board's inability to be stable, possibly frying your mobo, above 2.2v. This is the main reason why I said to stay below 1000mhz. Just an advice and of course YMMV.
Click to expand...

On the reference boards I had no problem up to 1200MHz on any of my boards. On the non-reference boards (P5N32-E SLI, Striker Extreme) I couldn't run past 900MHz on any of them.

hdnut said:
http://www.evga.com/forums/tm.asp?m=78381

If you read the post, it tells you how many mosfets this particular 680i has. You can also count them yourself in the pics provided.

http://www.evga.com/forums/tm.asp?m=123659&mpage=1&key=mosfet
Click to expand...

I never bothered to count them before. Some of my boards like my Black Pearl have the larger amount of them. (I'll confirm this later, but I'm fairly certain they do.)

hdnut said:
Many have ran 65nm with success and with decent overclocks on these 680i's. The main problem are the 45nm quads which can't be run worth a monkey's doo-doo.
Click to expand...

Yeah, that's unfortunate business.

hdnut said:
Since owning my 680i since March of 2007, I had never had to RMA it once. I have brought it back from what I thought at first was death or fried memory modules (crap load of "C1" errors or "--" or "D1" and bunch of others). Neither was true. The board is, however, very picky with the type of memory you use. To OP, I highly recommend sticking with ddr2 800, preferably micron d9's, and OC'ing these memory to much higher frequencies. Reason I state this is because the only time I had issues with memory was due to garbage EPP settings that the board did not like, which would either give off the "C1" code and refusing to boot or automatically set the voltages way too high.
Click to expand...

Ahhh the C1 errors. I didn't encounter those too often but on the ASUS Striker Extreme I got the CPUINIT error more times than I could count.
 
Dan_D said:
To be fair, I overclocked the crap out of them using dual and quad core processors. The systems were perfectly solid and stable from anywhere between 3 days to 3 months. After that the boards all died. Many of the dead 680i SLI boards I've had were DOA. Almost half of them were. The other half all died on me. One of them was killed by a watercooling leak, but other than that one board the rest have no excuses.

The 680i SLI still doesn't overclock quad core processors as well as other chipsets do. The 680i SLI still has voltage regulation issues and BIOS issues causing any number of issues. The most noteworthy problem is that the 680i SLI reference boards rarely work correctly with storage controllers thanks to spread spectrum settings changes in the BIOS. Let's also not get started on memory compatibility problems and lack of 45nm quad core support.

The 780i SLI is the same MCP and SPP that the 680i SLI boards used. The only difference is an added third chip that adds PCI-Express 2.0 support and 8 additional PCI-Express lanes. While these boards SHOULD have been better, they weren't. Almost none of the 680i SLI's problems were fixed on them. The 790i SLI Ultra is better, but it has data corruption issues. Though to be fair the 680i SLI and 780i SLI problems I have personally experienced and verified, while the 790i Ultra SLI problems I have not seen for myself. My experiences with the 790i Ultra SLI were actually pretty good.

Today with ATI's cards being the better deal in terms of price vs. performance and with the 4870 X2's release drawing near there is NO compelling reason in my mind to go with an NVIDIA chipset based board.
Click to expand...


I guess I need the editor tag over my name for people to listen to me... I said the same things all over this forum (nvidia has flaky chipsets) and I get blasted for it... oh well...
 
screwballl said:
I guess I need the editor tag over my name for people to listen to me... I said the same things all over this forum (nvidia has flaky chipsets) and I get blasted for it... oh well...
Click to expand...

You've been here for three days and you're very outspoken. I don't see why you're so surprised people aren't quick to agree with you.
 
screwballl said:
I guess I need the editor tag over my name for people to listen to me... I said the same things all over this forum (nvidia has flaky chipsets) and I get blasted for it... oh well...
Click to expand...

There is also a difference between saying "OMG NVIDIA chipsez are teh suxors!!!!!" and providing information concerning the chipset and its' flaws. I've provided information that can be found all over the web as well as reproducable experiences with those chipsets that substantiate my opinion concerning those chipsets.
 
Dan_D said:
There is also a difference between saying "OMG NVIDIA chipsez are teh suxors!!!!!" and providing information concerning the chipset and its' flaws. I've provided information that can be found all over the web as well as reproducable experiences with those chipsets that substantiate my opinion concerning those chipsets.
Click to expand...

I agree.

This is what i think:
I personally love my 680i SLI chipset it works fine. But when people just say how it sucks all the time with out stating reasons why, gets pretty annoying. And they need to realize that not everyone has had millions of them die on them, even if you or anyone else consider them flaky.

And remember that this are peoples opinions, you dont know what they had been doing to that board. My 680i is running everything stock and people have told me how much it sucks because they cant OC or they never POST because the board wont post with there OCed RAM modules. So i dont really worry much about the chipset more about the people using it and what they were trying to do. Now, there are those 680i's that are non-user error(like wont post at stock clocks or the data corruption issue, but in that case i just say rma and suck it up).

It just depends what you want to do and what the chipset is known for doing about it. If you OC and you know for a fact the 680i's are more likely going to be unstable then don't buy one(this doesn't count for the people at launch since it was a new chipset at the time). As for things like data corruption all you can really do is suck it up and hope you get a good one or if theres a BIOS fix that can be downloaded, try that out.(or rma). The 680i chipset is not all bad, but normally people who try and do ''extreme'' things to it have a problem, then some go on forums and say how much is sucks a$$ because they couldn't get it up to insane numbers.

As for the data corruptions and POST errors and other non-user faults, all that can be said is that they aren't all like that and you not need to hope you get one that works out of the box. But if you somehow get error with all your boards out of the box then go ahead and swich to a different chipset and post your opinion is a more calm manner and realize that there are people out there that enjoy the chipset and have yet to have a problem.

All i can say is that the 680i is a nice stable chipset at stock clocks for me. But i am not going to go around saying that teh 680i is teh best chipsets everr!!11, because i know it has faults, but me and tons of other people are enjoying are the chipset(OCed or not) even if others think there all flawed.

Just my 2 cents.;)
 
Yeah, I would avoid a 680i at this stage. If you don't plan on SLI it's really more trouble than it's worth. In fact with ATI doing so well with their cards right now I'm going to switch back to an Intel chipset.
 
convexion said:
Yeah, I would avoid a 680i at this stage. If you don't plan on SLI it's really more trouble than it's worth. In fact with ATI doing so well with their cards right now I'm going to switch back to an Intel chipset.
Click to expand...

i would say the same, but a P35 chipset is worthless for crossfire, so i'd have to buy an X38, P45, or X48 chipset. which would all cost me everything i have plus an arm or leg. so that isn't happening :/
 
convexion said:
Yeah, I would avoid a 680i at this stage. If you don't plan on SLI it's really more trouble than it's worth. In fact with ATI doing so well with their cards right now I'm going to switch back to an Intel chipset.
Click to expand...


Yeo, but in a little while Nvidia will have the upper hand.
 
3far4shot said:
Yeo, but in a little while Nvidia will have the upper hand.
Click to expand...

We don't know that for sure. Last time NVIDIA really had the upper hand it took almost two years for NVIDIA to break even with ATI.
 
Dan_D said:
We don't know that for sure. Last time NVIDIA really had the upper hand it took almost two years for NVIDIA to break even with ATI.
Click to expand...

They should make a GTX 280 X2, that could combat the 4870 X2, but knowing Nvidia it would probably cost around $800 and up.:( Which is the problem.
 
3far4shot said:
They should make a GTX 280 X2, that could combat the 4870 X2, but knowing Nvidia it would probably cost around $800 and up.:( Which is the problem.
Click to expand...

I don't think that a GX2 280 or something like that will happen until NVIDIA does a die shrink of the GPU.
 
AmongTheChosenX said:
yeah. and thats WHEN that actually happens. lol
Click to expand...

Who knows? I think that a die shrink and a reduction in power consumption and the thermal envelope of the GPUs will be required to put two of those things on one card. They may do the sandwiched PCB again ala Geforce 7950GX2/Geforce 9800GX2 but either way they do it they'll have to overcome the power and heat issues.
 
You must log in or register to reply here.
Back
Top