Graphics card power usage anomaly

How can the R9 Nano use less electricity than the R9 280X when it's more powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

Surely the GPU uses way more than the RAM?

And since it's the same nm process, the individual transistors must use the same amount of power each. And there are twice as many of them.

On 8/7/2020 11:01 AM, Commander Kinsey wrote:
On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

Surely the GPU uses way more than the RAM?

And since it's the same nm process, the individual transistors must use
the same amount of power each.* And there are twice as many of them.

Blow it out your ass, pasty face!

On Sat, 08 Aug 2020 00:53:35 +0100, Ricketty C wrote:

On Friday, August 7, 2020 at 7:02:26 PM UTC-4, Jeff Hickling wrote:
On 8/7/2020 11:01 AM, Commander Kinsey wrote:
On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

Surely the GPU uses way more than the RAM?

And since it's the same nm process, the individual transistors must use
the same amount of power each. And there are twice as many of them.

Blow it out your ass, pasty face!

Wow! I realize some here don't have much patience with people who think they know things they don't, but that seems a bit over the top!

I agree that Commander Kinky seems to be a bit rude. He asks a question and when someone replies he seems a bit less than appreciative.

But it's not like he insulted anyone or impugned their integrity. What happened to a response in kind?

Hey asshole!! Don't **** about with the group headers!

In article op.0ozq0koqwdg98l@glass,
Commander Kinsey wrote:

Surely the GPU uses way more than the RAM?

That will depend on the actual usage at the time. In a lot of modern
designs, whole blocks are designed to clock-stop themselves (partially
or completely) if they aren't active. This cuts their power usage
down a great deal.

And since it's the same nm process, the individual transistors must use the same amount of
power each. And there are twice as many of them.

That's a non sequitur, I believe... your mental model of how
transistors use power is a bit too simple.

Modern CMOS gates draw a small amount of "static" power when they're
not changing state (basically, a leakage current), and a great deal
more power when they're changing from 0 to 1 or 1 to 0. The power
usage thus depends both on the process (smaller-nm processes tend to
have higher leakage currents _unless_ you drop the voltage) and on the
amount of activity.

So, if you throw more transistors into a device, but run them slower
(or don't clock more of them on average than you did in a smaller
device) you can end up with less dynamic power usage.

Slowing the RAM speed down by 3:1 is going to result in a big decrease
in power usage.

If you use the same-nanometer process size, but run it at a lower
voltage, you reduce the static (leakage) power usage.

If you have a different DRAM which can hold charge in its cells
longer, you can run it at a slower refresh rate, and save power that
way.

Commander Kinsey wrote:
On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

Surely the GPU uses way more than the RAM?

Why "surely"?


And since it's the same nm process, the individual transistors must use
the same amount of power each. And there are twice as many of them.

For the same architecture, sure. They aren't so all bets are off.

On Sat, 08 Aug 2020 02:09:11 +0100, Dave Platt wrote:

In article op.0ozq0koqwdg98l@glass,
Commander Kinsey wrote:

Surely the GPU uses way more than the RAM?

That will depend on the actual usage at the time. In a lot of modern
designs, whole blocks are designed to clock-stop themselves (partially
or completely) if they aren't active. This cuts their power usage
down a great deal.

However I was comparing maximum power usage and maximum computational speed.

And since it's the same nm process, the individual transistors must use the same amount of
power each. And there are twice as many of them.

That's a non sequitur, I believe... your mental model of how
transistors use power is a bit too simple.

Modern CMOS gates draw a small amount of "static" power when they're
not changing state (basically, a leakage current), and a great deal
more power when they're changing from 0 to 1 or 1 to 0. The power
usage thus depends both on the process (smaller-nm processes tend to
have higher leakage currents _unless_ you drop the voltage) and on the
amount of activity.

So, if you throw more transistors into a device, but run them slower
(or don't clock more of them on average than you did in a smaller
device) you can end up with less dynamic power usage.

Slowing the RAM speed down by 3:1 is going to result in a big decrease
in power usage.

If you use the same-nanometer process size, but run it at a lower
voltage, you reduce the static (leakage) power usage.

If you have a different DRAM which can hold charge in its cells
longer, you can run it at a slower refresh rate, and save power that
way.

I see, thanks for the detailed answer.

On Sat, 08 Aug 2020 17:19:41 +0100, Chris wrote:

Commander Kinsey wrote:
On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

Surely the GPU uses way more than the RAM?

Why "surely"?

I was thinking of the colossal heatsink attached to it. But then I suppose the RAM does actually contact that same heatsink. Although on a card I just dismantled, the RAM chips are tiny compared to the GPU, I very much doubt half the power could come from those tiny things.

And since it's the same nm process, the individual transistors must use
the same amount of power each. And there are twice as many of them.

For the same architecture, sure. They aren't so all bets are off.

On Sat, 08 Aug 2020 02:52:03 +0100, Ricketty C wrote:

On Friday, August 7, 2020 at 8:20:49 PM UTC-4, Tabby wrote:
On Saturday, 8 August 2020 00:53:39 UTC+1, Ricketty C wrote:
On Friday, August 7, 2020 at 7:02:26 PM UTC-4, Jeff Hickling wrote:
On 8/7/2020 11:01 AM, Commander Kinsey wrote:
On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

Surely the GPU uses way more than the RAM?

And since it's the same nm process, the individual transistors must use
the same amount of power each. And there are twice as many of them.

Blow it out your ass, pasty face!

Wow! I realize some here don't have much patience with people who think they know things they don't, but that seems a bit over the top!

I agree that Commander Kinky seems to be a bit rude. He asks a question and when someone replies he seems a bit less than appreciative.

But it's not like he insulted anyone or impugned their integrity. What happened to a response in kind?

he's a time wasting troll.

Yeah I guess you are right about that. What is he talking about my munging the "group headers"??? Is he being delusional?

Because you deleted the crosspost. The discussion is taking place between people in more than one group, and you deliberately cut half of those people out.

On Sat, 08 Aug 2020 21:15:03 +0100, Ricketty C wrote:

On Saturday, August 8, 2020 at 3:08:17 PM UTC-4, Commander Kinsey wrote:
On Sat, 08 Aug 2020 02:52:03 +0100, Ricketty C wrote:

On Friday, August 7, 2020 at 8:20:49 PM UTC-4, Tabby wrote:
On Saturday, 8 August 2020 00:53:39 UTC+1, Ricketty C wrote:
On Friday, August 7, 2020 at 7:02:26 PM UTC-4, Jeff Hickling wrote:
On 8/7/2020 11:01 AM, Commander Kinsey wrote:
On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

Surely the GPU uses way more than the RAM?

And since it's the same nm process, the individual transistors must use
the same amount of power each. And there are twice as many of them.

Blow it out your ass, pasty face!

Wow! I realize some here don't have much patience with people who think they know things they don't, but that seems a bit over the top!

I agree that Commander Kinky seems to be a bit rude. He asks a question and when someone replies he seems a bit less than appreciative.

But it's not like he insulted anyone or impugned their integrity. What happened to a response in kind?

he's a time wasting troll.

Yeah I guess you are right about that. What is he talking about my munging the "group headers"??? Is he being delusional?

Because you deleted the crosspost. The discussion is taking place between people in more than one group, and you deliberately cut half of those people out.

You fail to understand that many people here are on google groups and don't have the ability to post to multiple groups. So that's on you.

What the **** are people doing using that web based ****? I suppose you read your emails on a webpage too? Get a newsreader, they're free!

Commander Kinsey wrote:
On Sat, 08 Aug 2020 17:19:41 +0100, Chris wrote:

Commander Kinsey wrote:
On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's
more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one
third the
speed. Simples.

Surely the GPU uses way more than the RAM?

Why "surely"?

I was thinking of the colossal heatsink attached to it. But then I
suppose the RAM does actually contact that same heatsink. Although on a
card I just dismantled, the RAM chips are tiny compared to the GPU, I
very much doubt half the power could come from those tiny things.

And since it's the same nm process, the individual transistors must use
the same amount of power each. And there are twice as many of them.

For the same architecture, sure. They aren't so all bets are off.

Well, the RAM chips on the Nano are very tiny.

R9 280X 60% 250W $299 4.3Bil 28nm 2048/128/32 GCN 1.0 GDDR5 288GB/s
R9 Nano 100% 175W $649 8.9Bil 28nm 4096/256/64 GCN 3.0 HBM 512GB/s

The Nano uses HBM, which is mounted inside the GPU. There's
no ring of chips around the outside of the GPU on the Nano.
One of the tricks, is getting the GPU and HBM to the same
height, so that the cooler meets them properly. This is easiest
to do, if there is only one supplier of HBM. There might have
been a problem in a later HBM product, with two suppliers of
HBM die, who were not making the chips the same height. There
was a fair bit of whinging about how the cooler was supposed
to fit on the later stuff. If you're taking an HBM product
apart, you'd need to "read the ancient texts" to see if there
are any rules to follow during disassembly.

I'm not convinced all the Wiki info I'm seeing is correct. It's
possible the process the 28nm is done in, is actually different,
and there could be a leakage mitigation method in place on
the newer one.

There is too much of a difference between them, to just
hand-wave it away.

If you have a newer card, the chip power management is dynamic,
and a function of what the card is doing. The cards can go
into clock-limited mode (Vcore to max to allow the clocking,
power is much less than TDP), or TDP-limited mode (TDP power
limiter trips, clock rate is reduced to stay within TDP).

Perfcap: "VREF" = Vcore_Max, power is 1/3rd of allowed TDP

https://i.postimg.cc/GhvnCqFw/Smoke-Particles2.jpg

Perfcap: "PWR" = TDP limit at full power,
Vcore is reduced to meet limit, Clock drops because of VCore

https://i.postimg.cc/85cZzPxf/furmark.jpg

And stuff like that makes hand-waving more difficult.

Paul

On 07/08/2020 19:01, Commander Kinsey wrote:
On Fri, 07 Aug 2020 13:31:09 +0100, Chris wrote:

Commander Kinsey wrote:
How can the R9 Nano use less electricity than the R9 280X when it's more
powerful with more transistors but the same nm process?
https://www.techpowerup.com/gpu-spec...-r9-nano.c2735
https://www.techpowerup.com/gpu-spec...-r9-280x.c2398


Completely different architecture and the RAM is running at one third the
speed. Simples.

Surely the GPU uses way more than the RAM?

And since it's the same nm process, the individual transistors must use
the same amount of power each.* And there are twice as many of them.

If it's designed more cleverly is could be that on average the
transistors spend less time doing work (and more time in a static state).

--
Brian Gregory (in England).

On 08/08/2020 21:26, Commander Kinsey wrote:
What the **** are people doing using that web based ****?* I suppose you
read your emails on a webpage too?* Get a newsreader, they're free!

Why would anyone take the advice of someone who doesn't even know about
how digital transistor circuits use power?

--
Brian Gregory (in England).

On 08/08/2020 20:08, Commander Kinsey wrote:
On Sat, 08 Aug 2020 02:52:03 +0100, Ricketty C
Yeah I guess you are right about that.* What is he talking about my
munging the "group headers"???* Is he being delusional?

Because you deleted the crosspost.* The discussion is taking place
between people in more than one group, and you deliberately cut half of
those people out.

You can't tell somebody off because the news server they connect to
doesn't happen to carry all the newsgroups you posted to. Idiot.

--
Brian Gregory (in England).