Thermal issue : Will EE3 be able to survive the heat?
- Thread starter DeadmeatGA
- Start date
randycat99
Veteran
Regarding an earlier post, I believe the 90W G5 rating was actually for a dually system. I could be wrong though.
Nope, went searching & found this http://www.aceshardware.com/forum?read=105035992. Not actually TDP as I'd remembered but it was 90W.
Don't forget smaller mircon process = smaller die = less surface area for cooling. Thats why amd has the heatspreader on the athlon 64s .
Heat is going to be a problem with all the next gen consoles.
Thats why in the xbox 2 i'm expecting the cpu to be 500-1ghz slower than the top of the line cpus of the time. That way if a 5ghz p4 has a voltage of 1.7 they can run a 4ghz at 1.5 or 1.4 . It would also put off less heat because of the decrease in speed = a decrease in power consumption. Since they are going with ati for the gpu I don't see heat as an issue. The chips all run cool compared to other chips in gpus.
Heat is going to be a problem with all the next gen consoles.
Thats why in the xbox 2 i'm expecting the cpu to be 500-1ghz slower than the top of the line cpus of the time. That way if a 5ghz p4 has a voltage of 1.7 they can run a 4ghz at 1.5 or 1.4 . It would also put off less heat because of the decrease in speed = a decrease in power consumption. Since they are going with ati for the gpu I don't see heat as an issue. The chips all run cool compared to other chips in gpus.
After 20 topics a week of bashing and doubts I am now convinced that Deadmeat is right. PS3 will overheat after one hour of playing if they don't lower their ambitions to just 60 gigaflops. Anyway, PS3 will suck really bad and Xbox2 will blow Kutaragi's hardware out of the water. He just gave us too much proof to be denied and that's just the way it's going to be. Thanks Deadmeat.
Fafalada said:
That's just because "derogitory" is his default state concerning all things Sony (at least all things Playstation) and he always finds a way to stick in a snipe somewhere (for instance Officially no. Unofficially, the PSX existed ever since Nintendo dumped the PS(PlayStation) and Kutaragi swore his vengence.), which is a shame because he CAN post things that are perfectly valid subjects to discuss, but it's hard to when there's at least one troll (or doomsaying inference) per post. So tiresome... -_-
Regarding the heat itself, we of course don't have enough information to know just what it will produce, but I imagine they'll be spending more time investigating many other cooling methods than just "fan in da back."
I'm sure they won't do anything like Apple's method with the G5, but I rather like their design.
Panajev2001a
Veteran
Look at that and then look here:
http://appft1.uspto.gov/netacgi/Par...&s1=20030037221&OS=20030037221&RS=20030037221
That architecture IS being designed with power saving in mind.
Example:
scalar execution causes the FXU/FPU unitilized to sit idle and go in sleep mode, if several APUs are working together for a task, before the synchronization point the ones that have finished do go on stabdy and consume less power.
Another small example: the Register file is also designed to be simplier and save on power: knowing that FXU and FPUs cannot access in parallel the register file, you do not need to add tons of ports or double clock it.
Not only the Banias can be designed to turn off unusued units in cases like branch misprediction or similar situations which would cuase parts of the CPU to sit idle.
Most modern cpus have features like that , the athlons and the p3s have the halt command , p4s have clock throttling . I'm not say the chip will not hit 3ghz cause it can. But i believe at some point heat will become a huge factor. Now if its at 1ghz , 2 ghz , 3ghz , 4 ghz or 5ghz it will become an issue no matter how much powersaving features a chip has.
Panajev2001a
Veteran
JVD, this goes beyond thermal clocking of Pentium 4: I doubt that in the actual Pentium 4 that if the SSE units are not used and the FPU is not used the CPU turns them off ( on a cycle basis ), something similar could happen in their mobile chips, but not too efficiently.
Ar you tell me that when the SSE unit executes a serial instuction that the other parallel execution units are shut off ?
There is a way to do things better and to have great performance: look at the Banias, not even designed by the American guys at Intel U.S.A. and it is a performer AND a chip with very LOW power consumtpion.
Also who said that leakage with STI's 45-65 nm processes will be as bad as it was for several 90 nm processes: they are developing the new manufacturing processes with CELL in mind afterall ) and they are investing quite a lot of money in this area as well ).
Ar you tell me that when the SSE unit executes a serial instuction that the other parallel execution units are shut off ?
There is a way to do things better and to have great performance: look at the Banias, not even designed by the American guys at Intel U.S.A. and it is a performer AND a chip with very LOW power consumtpion.
Also who said that leakage with STI's 45-65 nm processes will be as bad as it was for several 90 nm processes: they are developing the new manufacturing processes with CELL in mind afterall ) and they are investing quite a lot of money in this area as well ).
If you have an 86 watt cpu you need double that for a peltier to cool it. But on the flip side the hot side of the peltier would be twice as hot as the cpu was to begin with. So you'd need an even better heatsink /fan or waterblock / radatior to cool the backside of the peliter. That is why its only used in very high end water pcs.PC-Engine said:
DeadmeatGA
Banned
...
The traditional PC type power management tricks don't work well on a game console CPU. PC processors have the option of turning off idle units and slowing down the clock rate because most applications are not specifically optimized for it and have a variable processing load, giving the CPU a breathing room to rest and cool down.
On the other hand, the console games are written for and specifically optimized to maximize the usage of target hardware and don't allow the CPU to cool down. What do you think is going to happen when the user plays GT6 that has 90~95% CPU utilization on PSX3 three hours straight???
The only real solution is to physically lower the clockrate cap so that the EE3 doesn't overheat under the worst circumstances.
The traditional PC type power management tricks don't work well on a game console CPU. PC processors have the option of turning off idle units and slowing down the clock rate because most applications are not specifically optimized for it and have a variable processing load, giving the CPU a breathing room to rest and cool down.
On the other hand, the console games are written for and specifically optimized to maximize the usage of target hardware and don't allow the CPU to cool down. What do you think is going to happen when the user plays GT6 that has 90~95% CPU utilization on PSX3 three hours straight???
The only real solution is to physically lower the clockrate cap so that the EE3 doesn't overheat under the worst circumstances.
DeadmeatGA
Banned
...
Let's study Xscale as a power requirement example.
600 Mhz : 0.5 Watts
1000 Mha : 1.5 Watts
The power usage trippled, even though the clock rate increated by 66%. There is almost an quadratic relationship between the clock rate and power consumption. If you clock EE3 at 3 Ghz, then it will burn 80~130 Watts just like any other gigahertz CPUs, and worse yet, it has to sustain that clockrate under heavy processing load for a prolonged period.(GT6) Can the EE3 really take that much abuse? I am not sure. Eventually, SCEI might need to permanantly cap the EE3 clockrate to limit the power consumption to 50 watts or less just to prevent the melt-down.
Let's study Xscale as a power requirement example.
600 Mhz : 0.5 Watts
1000 Mha : 1.5 Watts
The power usage trippled, even though the clock rate increated by 66%. There is almost an quadratic relationship between the clock rate and power consumption. If you clock EE3 at 3 Ghz, then it will burn 80~130 Watts just like any other gigahertz CPUs, and worse yet, it has to sustain that clockrate under heavy processing load for a prolonged period.(GT6) Can the EE3 really take that much abuse? I am not sure. Eventually, SCEI might need to permanantly cap the EE3 clockrate to limit the power consumption to 50 watts or less just to prevent the melt-down.
I agree. This is IMO the biggest issue of all next-gen consoles, especially since the future processes are so much more leaky. Not only do consoles need better heat management, they also need to produce less heat in general. This is why the current batch consoles have so low specs. A 733 P3 for the Xbox, only 32-64MB of ram, etc. They have smaller sizes vs. PCs and will be less carefully used.
The effect will be that the future consoles will be big and heavy, bigger than the Xbox now, so as to guard against overheat and to fit big HSFs. Clock speeds will be very modest, as much as 2/3 or 1/2 of the fastest CPUs and GPUs available. I don't expect the PS3 CPU (BE) to do better than CPUs do now in clockspeed. Probably <2Ghz for the CPU and around 1Ghz for the GPU.
The effect will be that the future consoles will be big and heavy, bigger than the Xbox now, so as to guard against overheat and to fit big HSFs. Clock speeds will be very modest, as much as 2/3 or 1/2 of the fastest CPUs and GPUs available. I don't expect the PS3 CPU (BE) to do better than CPUs do now in clockspeed. Probably <2Ghz for the CPU and around 1Ghz for the GPU.
DeadmeatGA
Banned
...
Exactly. The CELL processor described in Suzuoki patent application presumes four CELL cores running at 4 Ghz, which is impractical for consoles. A server chip might do it with exotic cooling scheme, but a console chip??? Forget it. What you saw in that patent application is not the EE3, but certain unnamed high-performance server chip intended for Sony's high-performance graphics server/workstation business. The EE3 will be much more humble.
zidane1strife
Banned
65nm... the speed estimates... I've heard... surpass all current... by an order of magnitude... delusional thoughts of engineers in the past? or....
I don't think the home consoles necessarily need to be that much bigger to rememdy heat problems etc. Just look at top of the line laptops which have ZERO free internal space. If we use that as reference then modify it slightly by doubling the internal case volume, it shouldn't be a problem to add in a high power GPU. Xbox didn't need to be the size that it is. IMO the Xbox 2 would be able to fit into a case that's 8"Wx3"Hx8"D.
Yea it's too bad SCEI made the patent and SCEI makes Playstations...
It's no workstation, Patent says 2 PE's would make a workstation.
Patent never says what FIGURE 6 IS, but it goes to great lengths to establish(hint) it as a game-console. Notice the diagram of it? I think it was the only embodiment to actually have a whole diagram.
Face the facts DM, SCEI is manufacturing "Cell" for use in PS3 as stated by the PR's and Sony themselves. This will take place in the 2004 time-frame.
Let's look at what Kutaragi and the others say what Cell is? 1TFLOPS, that's the only thing that they say "Cell" is. Kutaragi has made no mention of other different powered Cell chips other than it's 1TFLOPS base.
Why do you hate Sony THIS much? Are you afraid Sony will one-up MS next gen with it's ps3? So much afraid that you must bash Cell, ps3, psp and doom them to failure? Your no CMOS engineer, your not one of the 300 engineers that took part in the Cell project since 2000. Your nothing but a troll who has a dream, a dream to see Sony crash and burn.
Kutaragi had a point where he said that there were people who critisized other people's projects that were not any type of engineer but a common person. This is you, your not sony Toshiba or IBM, you have no right to doom things to failure.
Similar threads
