Quote:Yeah the static power was marginally improved with 22nm over 32nm, not too surprising considering that with the node shrink all the dielectric insulating layers got thinner.

Ordinarily we would have expected leakage to get worse, not better, but the Finfets really helped with that.

What really saved 22nm from a static leakage perspective is that Intel was able to scale-down the voltage needed to hit any given clockspeed. (clockspeed is not a factor in static leakage, only temperature and voltage are factors in static leakage)

Quote:Oh dear, please do not comment on topics, you are apparently no expert of. 
Let me first state that when idle, leakage is the single biggest contributor to power. In other use cases active power is (still) dominating. The relation between these to values however goes up with each smaller process node in favor of leakage. What you also should know is, that leakage not just doubles when moving down half a node. Leakage is one of the main effects working against Moores law.

The main reason to going to 3d transistors/FinFETs is to have more control of the geometric layout of the gate channel and the electric field controlling the channel, thus reducing leakage current significantly.
To give you some idea. 22nm FinFET has much lower leakage than 28nm planar. I estimate that 14 nm FinFET is about the same as 28nm planar for logic but already worse for SRAM. 14nm FinFET is in any case worse then 22nm FinFET, because you cannot defeat physics. 

At this point you need to think about clever designs, as for instance the big.little concept. Keep in mind Intel is already past the general FinFET/3-gate gain with respect to leakage.

Mind to enlighten me, how is this supposed to improve when going down to 10nm, even if you consider that you are using compound materials? I mean at some point you also need to reduce the thickness of the gate dielectric, where high-k materials are already used to extend Moores law.

(07-01-2016, 05:38 PM)RolloTheGreat Wrote: http://www.grapheine.com/bombaytv/graphiste-en-5e54fe59139f98fd601f6e21a063c899.html

(07-02-2016, 05:41 AM)SickBeast Wrote: This is looking really bad for AMD!  The RX480 cards are killing all sorts of motherboards!  It even looks like it killed apoppin's board! :(

(07-02-2016, 05:44 AM)SteelCrysis Wrote:

(07-02-2016, 05:41 AM)SickBeast Wrote: This is looking really bad for AMD!  The RX480 cards are killing all sorts of motherboards!  It even looks like it killed apoppin's board! :(

Really?

(07-02-2016, 07:28 AM)SickBeast Wrote: His board is dead.  He tried to fix it all day today.  He even bought a new PSU.  His computer won't boot.  He's buying a new motherboard this week.  Integrated graphics won't even work.  It looks like the RX480 may have completely killed his board.

Quote:We skipped long-term overclocking and overvolting tests, since the Radeon RX 480’s power consumption through the PCIe slot jumped to an average of 100W, peaking at 200W. We just didn’t want to do that to our test platform

(07-02-2016, 08:50 PM)BoFox Wrote: Poppin tried to overclock it sky-high?    

Tomshardware warned us:

Quote:We skipped long-term overclocking and overvolting tests, since the Radeon RX 480’s power consumption through the PCIe slot jumped to an average of 100W, peaking at 200W. We just didn’t want to do that to our test platform

(07-02-2016, 10:26 PM)SickBeast Wrote: I think what has happened is that AMD overclocked the RX480 at the last minute and did not test the card properly.  I say this for a few reasons:

1. The card has virtually no overclocking headroom.
2. The card runs quite hot even at stock settings.
3. The card draws more power than it's rated for.
4. The GTX 1080 and GTX 1070 came in way ahead of expectations in terms of price and performance.

I think AMD got really desperate and tried everything they could to make the RX480 competitive.  Unfortunately they failed badly. :(

(07-02-2016, 10:26 PM)SickBeast Wrote: I think what has happened is that AMD overclocked the RX480 at the last minute and did not test the card properly.  I say this for a few reasons:

1. The card has virtually no overclocking headroom.
2. The card runs quite hot even at stock settings.
3. The card draws more power than it's rated for.
4. The GTX 1080 and GTX 1070 came in way ahead of expectations in terms of price and performance.

I think AMD got really desperate and tried everything they could to make the RX480 competitive.  Unfortunately they failed badly. :(

(07-03-2016, 12:35 AM)SteelCrysis Wrote: AMD is working on a solution to the high power consumption: http://www.tomshardware.com/news/amd-power-gpu-rx-480,32200.html

(07-03-2016, 04:51 AM)Mousemonkey Wrote:

Quote:It's not bad, really - at just over half of the power consumption

(07-02-2016, 09:07 PM)BoFox Wrote: Yo Ocre, I've been thinking that AMD didn't get to completely solve the power consumption issue on GloFo's new 14nm process.  If AMD didn't diverge from TSMC and stuck with the same 16nm process as NV, I think that AMD had their RX 480 ready for launch maybe 6 months ago...  with lower idle power consumption than that of Tonga R9 285 (and lower load consumption than that of Pitcairn XT R9 270X -- less than 120W on average).  See how great the new Pascal cards are at idle consumption:




Could it also be that 8GB GDDR5 clocked at 8GHz effective would consume about 5W more than 7GHz that AMD is apparently setting lots of these 380 cards at default (not just for 4GB but 8GB cards as well)?  Furthermore, how much would 8GB consume over 4GB?  My guess is around 15W.

(07-03-2016, 06:10 AM)Mousemonkey Wrote: When the first reviews surfaced I skimmed over them and came to the conclusion that the card was pants and then this power issue came to light and I giggled so hard I nearly had a trouser accident.

That picture was from another THG Mod BTW but it did make me laugh.

(07-03-2016, 04:29 PM)ocre Wrote: (a bunch of stuff that is likely very true) 

(07-04-2016, 12:54 AM)ocre Wrote: Amd raja....need his head checked.  All that talking up he did (while talking down pascal) what a freaking joke.  

And this part is especially, i want bofox to take note of.
Take a look at the 380x the full tonga to which I believe that Fiji and the 480 come almost directly from.  The 480 is running about 30% higher clock speed and has 15% added compute units.  The 480 is not anywhere I would expect it to be given l the hype raja went on and on about, how they have created a radical newly improved design.  It's just not.  The tweaks they did to the arch were almost no more beneficial, It's just truly pathetic.

I am making such comparison to contrast all those saying pascal was reduced ipc. There are some stark differences, one being that the gp104 has a smaller bus, less bandwidth, far less rops, far less cores than the 980ti...it's logical that it would have some reduction in IPC at the same clock speed as the 980ti.  It's a much more stripped down and lean chip.

The 480 vs 380x- more compute units, higher bandwidth, higher speed, more memory....it's just not a great result.


Just clock speed increase alone, 380x times 130% and it's at 480 performance.  This is not taking in account the 15% extra compute units..which are also running 30% faster. It's just unbelievable, where is this per core improvement amd has chirped about?  Well, interestingly enough, they claim that the 480 cores are 15% "faster" than those on the r9 290.  Haven't crunched those figures yet but since the 380x has the same bus size, I didn't want to skew the results unfairly using hawaii's large bus against the 480.  The 380x is, for all intents and purposes, a pretty direct point of reference. The added cores of the 480 put the 380x at the disadvantage yet clock speed alone would account for almost all of the improvement we see.

It's very very interesting.  There has to be some reason these "improvements" amounted to very little.  Amd talks about many of them and have graphs and claims of huge gains...but they are not there in the real world.  They are simply not adding up.  How?  Amd lies with hyper exaggerated figures, compulsively....or they are really lacking on producing outcome from some of these good ideas.  Relying mostly on full computer Automated chip design would surely hold back, especially in finet, which I hear has to be worked st painstakingly with engineers piecing together like a jig saw puzzle in order to extract and harness the technology fully

(07-03-2016, 12:33 PM)BenSkywalker Wrote:

Quote:It's not bad, really - at just over half of the power consumption

Its' performance per watt is almost dead even with a Titan X.

Massive 28nm die, Budget 14nmFF die......

You are correct, it isn't bad, it is an embarrassment to bipeds.

(08-06-2016, 05:32 AM)SteelCrysis Wrote: http://techreport.com/news/30482/report-amd-shipping-100k-more-polaris-10-gpus-to-board-partners

(08-06-2016, 04:45 PM)RolloTheGreat Wrote:

(08-06-2016, 05:32 AM)SteelCrysis Wrote: http://techreport.com/news/30482/report-amd-shipping-100k-more-polaris-10-gpus-to-board-partners

I have no clue whatsoever why people are buying these.

They don't offer more performance than the cards they already have, so what is thee point?

"Yay! Yay! I bought a new card with the same performance I had 3 years ago! Yay! My computer stays the same! I would be confused by more performance!"

(08-07-2016, 09:44 AM)SickBeast Wrote:

(08-06-2016, 04:45 PM)RolloTheGreat Wrote:

(08-06-2016, 05:32 AM)SteelCrysis Wrote: http://techreport.com/news/30482/report-amd-shipping-100k-more-polaris-10-gpus-to-board-partners

I have no clue whatsoever why people are buying these.

They don't offer more performance than the cards they already have, so what is thee point?

"Yay! Yay! I bought a new card with the same performance I had 3 years ago! Yay! My computer stays the same! I would be confused by more performance!"

Kind of like when you bought an Xbox One when you already had a PS4?  What did you say?  "Yay!  Yay!  I bought a new console with less performance than what I had a year ago!  Yay!  My console is now less powerful!"