NVIDIA GeForce 4XX Series Discussion

[GSquadron]

Unfortunately it has 2x1 GB hynix ram. I think that CEO MSI 5870 lightning edition might be with the new chips and as reviews say the lightning editions like the LN2 (liquid nitrogen)
http://www.fudzilla.com/content/view/16889/34/
But the ram chips dont do anything new to the 1GB cards cuz they take the same place in the pcb of the card. It matters a lot for the 5970.

 

[Benetanegia]

448 is still a good bit less than 512, still wonder how performance will be. I too agree with it having to be yield problems, what other logical reason could there be to lower the specifications.

The only reason is power envelope. Exceeeding some certain envelope in a desktop card doesn't matter at all. Exceeding it even by 5w in a card that is going to be used in the dozens, hundreds or even thousands of them matters a lot. Like I said 12 ram chips consume at least 30w on average and that's a lot if you think about it. Anyway, servers usually have very limited power supplies for all the things they are powering.

 

[ToTTenTranz]

Fermi is too advanced for the current 40nm manufacturing.... it should have been made with 32nm

lol yeah you could call it "too advanced".. or you could also call it "too big".

Both can be used to describe that GPU.

 

[dir_d]

lol yeah you could call it "too advanced".. or you could also call it "too big".

Both can be used to describe that GPU.

LOL i was gonna say the same thing...Nvidia had to answer or they woulda been outta the game for a year or maybe more if they went straight to 32nm but you best believe there will be 32nm refreshes. NV already said this Fermi architecture will be around for 4 or 5 years i expect revisions and renames like the G80 days.

 

[GSquadron]

They better face the truth and show us what nvidia got, better or worse than ati. This move of nvidia was worse than i thought

 

[dir_d]

They better face the truth and show us what nvidia got, better or worse than ati. This move of nvidia was worse than i thought

Well im pretty sure it will be better but for them to be successful they need to play ati's game which is price/performance. If they drop fermi at 400 to 500 and its 20% faster it will be good for NV. They will still have problems though because the majority of the money to be made is in the mainstream market which ATI has locked down right now and NV is launching mainstream chips in the 2nd Quarter 2010.

 

[erocker]

The only reason is power envelope. Exceeeding some certain envelope in a desktop card doesn't matter at all. Exceeding it even by 5w in a card that is going to be used in the dozens, hundreds or even thousands of them matters a lot. Like I said 12 ram chips consume at least 30w on average and that's a lot if you think about it. Anyway, servers usually have very limited power supplies for all the things they are powering.

This is what I don't get. Why is 300w the magical number? Surely any enthusiast wouldn't mind something over 300w. If Fermi was a 400 watt monster that ate games for breakfast and worked really well, I'd be all over it!

 

[Solaris17]

This is what I don't get. Why is 300w the magical number? Surely any enthusiast wouldn't mind something over 300w. If Fermi was a 400 watt monster that ate games for breakfast and worked really well, I'd be all over it!

ya but thats a ludicrous amount of draw. i mean it would be awesome. but high power PSU's would be needed. especially if you wanted a multicard system youd need to upgrade more than just your current card.

 

[erocker]

ya but thats a ludicrous amount of draw. i mean it would be awesome. but high power PSU's would be needed. especially if you wanted a multicard system youd need to upgrade more than just your current card.

What so called "enthusiast" doesn't already have a high powered PSU? My system, overclocked and benching never draws over 480 watts. I would have no problems with a 3 or 4 PCI-E plug card. As long as it performs.

 

[Solaris17]

What so called "enthusiast" doesn't already have a high powered PSU? My system, overclocked and benching never draws over 480 watts. I would have no problems with a 3 or 4 PCI-E plug card. As long as it performs.

im not really arguing. just playing devils advocate i personally dont care. but their are people all around that have like a 5-600w psu and say im an enthusiest. which im not saying is wrong. but depending on how efficent fermi is and if it has power managment it could become a problem. especially if their PSU is old. not to mention the degridation that can come from loading a in-efficent 300w card often. that could wear a psu out.

 

[cadaveca]

What so called "enthusiast" doesn't already have a high powered PSU? My system, overclocked and benching never draws over 480 watts. I would have no problems with a 3 or 4 PCI-E plug card. As long as it performs.

Me either. If it does S3D @ 120FPS + Phys-X, I'm sold, no matter the power requirements. I'm sick of multiple cards, and related issues. Give it to me in one gpu, so I can put it in a MATX system, and save some desk space.

im not really arguing. just playing devils advocate i personally dont care. but their are people all around that have like a 5-600w psu and say im an enthusiest. which im not saying is wrong. but depending on how efficent fermi is and if it has power managment it could become a problem. especially if their PSU is old. not to mention the degridation that can come from loading a in-efficent 300w card often. that could wear a psu out.

Oh well..As far as I am concerned, cheaping out on a PSU for a high-end build is silly. but I guess you are partially right, as dirty/crappy power will probably lead to alot of card deaths. It's them wanting to sell as many as posisble that forces the constraints, really. Self-imposed.

I mean really, a 400W single card, or two 200W cards, does it matter?

 

[Bo_Fox]

It's for a workstation Tesla card.. those corporations that buy Tesla cards want reliable cards that will work with their server PSU's without creating too much heat. With 24 memory chips, NV had to make a tradeoff between those power-hungry GDDR5 chips and the power consumption of the GPU.

As for the enthusiast, I certainly hope that NV will be able to maintain 512 shaders with only 12 memory chips (1.5GB). As long as the idle power consumption is great, I do not care about the load consumption. Heck, there are some enthusiasts who have two 4870X2's or GTX 295's in their rigs, eating power like mad when under load. A single 4870X2 eats well over 300W by itself. I'd rather have a single GPU that's faster than a 5970X2 while eating less power.

If the GT300 can be cooled well enough with 512 shaders, then bring it on Nvidia, keeping in mind that a 5970X2 probably eats more power anyways. But wait a minute, NV has to do a dual-GPU card so NV does not want each of the dual-GPU chip to be THAT much worse than the single GT300.. oh well, that sucks!!!

 

[cadaveca]

If the GT300 can be cooled well enough with 512 shaders, then bring it on Nvidia, keeping in mind that a 5970X2 probably eats more power anyways. But wait a minute, NV has to do a dual-GPU card so NV does not want each of the dual-GPU chip to be THAT much worse than the single GT300.. oh well, that sucks!!!

Dual-gpu cards suck. It's so hard to cool gpu-sized silicon in such confined spaces...and given that we know they are having leakage issues, I certainly hope they do not do yet another sandwich-card. A good sandwich has the heat in the middle, not the bread! if I have no other choice than another sandwich for my performance "needs", I'll say no.

 

[Zubasa]

As for the enthusiast, I certainly hope that NV will be able to maintain 512 shaders with only 12 memory chips (1.5GB). As long as the idle power consumption is great, I do not care about the load consumption. Heck, there are some enthusiasts who have two 4870X2's or GTX 295's in their rigs, eating power like mad when under load. A single 4870X2 eats well over 300W by itself. I'd rather have a single GPU that's faster than a 5970X2 while eating less power.

If the GT300 can be cooled well enough with 512 shaders, then bring it on Nvidia, keeping in mind that a 5970X2 probably eats more power anyways. But wait a minute, NV has to do a dual-GPU card so NV does not want each of the dual-GPU chip to be THAT much worse than the single GT300.. oh well, that sucks!!!

IF you are talking about the "GTX 390", I guess it really isn't a problem that each GPU is weaker than their top-end single GPU cards.
As long as it is able to take a significant lead, it will be fine.

Also, the 4870X2 does not eat more than 300Ws.
The GTX 295 is more efficient, but both cards typically loads well under 300W and do not exceed their deisgned power.

 

[Benetanegia]

This is what I don't get. Why is 300w the magical number? Surely any enthusiast wouldn't mind something over 300w. If Fermi was a 400 watt monster that ate games for breakfast and worked really well, I'd be all over it!

First of all, the Tesla is rated below 225w. Now this is not about enthusiasts, Tesla is not for enthusiasts, it's for HPC a very different market where reliability and convenience matters much more than sheer power. In this particular case a TDP lower than (or around) 225w is almost a must, because most of the Tesla cards are probably going to replace 2 CPUs* in old servers, while almost the rest of the rig is going to be kept (especially the PSU and cooling system). By being limited to 225w a single Tesla can replace (the overall power consumption of) 2 typical server CPUs. For example, anyone can replace racks with an 8p server by a 4p + 2 Teslas, 2p +3 Teslas... whatever their requirements are, but they would not have to change anything (except the obvious). If power requirements were higher, they would start having problems finding a way to replace what they had for the cards and probably have to replace the entire powering + cooling system. Some Xeons and Opterons took quite long to be released at the same clocks as their desktop equivalents because of the same reason. Only when a revision made the TDP lower they were released.

You just can't compare the enthusiast market with the enterprise or HPC market (that's why Charlie is such a douche for the article he made), an enthusiast will willingly replace his PSU if it can't power something he really wants, in the HPC market that won't happen, basically because they would have to replace thousands of power supplies AND maybe even completely upgrade the power/cooling capabilities of the building or the floor where their server is, and that's not always posible. In fact, that's what marks the limit in performance for most supercomputers.

* Why 2 and not 1 or 3? Because it would mean the card would draw either 100-125w and would be too slow (or a miracle ) or 300w++ which would be way too much to cool on a single socket.

 

[rahulyo]

Nvidia Fermi

Is Nvidia really have any answer against Ati 5xxx series ???

 

[Shou Miko]

has any prices been released yet?

 

[KainXS]

well if you look at the prices of nvidia's high end launch prices then its somewhat like this in recent time, from what I remember the MSRP's for the cards.

7900GTX=$500
8800GTS640=400(but could be found for 350)
8800GTX=$600
GTX260=$400
GTX280=$650(was supposed to be 600)

do you see the trend

so based on that I would think that they would cost at least 500 dollars, more than likely 650 for the high end version, but guessing them is gonna be hard this time but just remember, nvidia hasn't made a flagship card for under 500 dollars in a while now.

 

[GSquadron]

looks like fermi is going to cost gtx395 for 700$!!!
Really stupid....
I hope prices don't go with that trend

 

[subhendu]

^^^ yea true...they will release 5 cards at first to kill Ati's five dx11 models

 

[shevanel]

i can see where alot of folks tax money is going to be spent..

 

[a_ump]

haha true, i had completely forgotten about taxes. can't wait to get mine. HD 5850 or a nvidia card if it performs and isn't outrageously priced...probly will be though XD

 

[erocker]

Two or three months to go yet. Maybe more?

 

[dir_d]

^^^ yea true...they will release 5 cards at first to kill Ati's five dx11 models

Nvidia said they are releasing GTX380 and 360 in Q1 2010 then in Q2010 the rest of the mainstream cards. This means you probably have to wait 6 months from now to get a mainstream Nvidia card.

 

[kid41212003]

Assassin's Creed II (PC) was delayed to Q1 2010 for DX11 support? (waiting for NVIDIA?)