RX Vega Multi-GPU Support is Here

FrgMstr

FrgMstr

Just Plain Mean
Staff member
2FA
Joined
May 18, 1997
Messages
58,687
New AMD Crimson drivers are out as of yesterday. Download links are on the release notes page. Here are the highlights. The most notable addition is that Vega finally gets CrossFire support, even though I guess we do not call it CrossFire any more? Or is this just true multi-GPU support for DX12 and VR? Hopefully one of you will know.

Radeon Software Crimson ReLive Edition 17.9.2 Highlights
Support For - Radeon RX Vega Series - Up to 2x Multi GPU support - Project CARS 2™ - Multi GPU profile support added
 
There is something to be said for refusing to brand a feature:
  1. You save on marketing costs.
  2. You mitigate the risk of "losing" the name brand battle.
  3. You don't have to worry about getting your branding on the board designer's box.
I applaud some of the moves AMD is making lately. They seem more focused on their product than they have been in a long time.
 
maxius said:
I shall assume the benchmarking is underway!
Click to expand...
It not worth cost of card at all from $550 to $800+ each.

OPS up date the look like the Vega 56 get a $50 price drop from nowinstock.net on newegg which still $100 over MSRP.
 
Last edited:
shspvr said:
That on apply to power not heat beside it doesn't cost that much at best maybe $3 to 5 more per month.
Click to expand...

I made no value propositions, just indicating "both are hot" is disingenuous because there's a really significant difference in power, especially with 2 cards as the article discusses.

And to clarify, power you draw to your computer does all end up as heat, unless you're charging a battery or similar to store/transfer the energy. The energy doesn't disappear.
 
PhaseNoise said:
I made no value propositions, just indicating "both are hot" is disingenuous because there's a really significant difference in power, especially with 2 cards as the article discusses.

And to clarify, power you draw to your computer does all end up as heat, unless you're charging a battery or similar to store/transfer the energy. The energy doesn't disappear.
Click to expand...
Ok I see
 
PhaseNoise said:
I made no value propositions, just indicating "both are hot" is disingenuous because there's a really significant difference in power, especially with 2 cards as the article discusses.

And to clarify, power you draw to your computer does all end up as heat, unless you're charging a battery or similar to store/transfer the energy. The energy doesn't disappear.
Click to expand...


This is not entirely true. I can promise you that while I may draw 500w at the wall, that my PC doesn't emit 500w of heat. Electricity is not energy, therefore the law of conservation of energy does not apply.
 
PhaseNoise said:
Could you please describe where the energy used from the wall is then?
Click to expand...

Energy is not used, it is converted.



If I turn a generator by hand, I have created emf, I have excited electrons by way of a rotating magnetic field to move along a conductor. If electricity was energy, then even when I stopped turning that generator, there would be stored potential in the windings of that generator via excited electrons. Electrons aren't energy, they are matter. The electrons are always there, whether excited or not. Magnetic flux is not energy. And a magnet is certainly not energy.
The energy you speak of is coal, steam, and heat back at the power plant.
There is a law all about conversion that states there is no 1:1, so if I draw 500w at the wall, it's impossible to dissipate 500w because the electrical energy is being used in other manners than a pure resistive load.

Now, if you are referring to electrical energy, then that is a different story.
 
Last edited:
alxlwson said:
Energy is not used, dude...
Click to expand...

Which is actually why my original posting explained where the energy went and concluded with the exact words "The energy doesn't disappear."

I am trying to figure out what you are saying happened to the energy.
 
PhaseNoise said:
Which is actually why my original posting explained where the energy went and concluded with the exact words "The energy doesn't disappear."

I am trying to figure out what you are saying happened to the energy.
Click to expand...


I already stated that electricity is not energy
 
alxlwson said:
This is not entirely true. I can promise you that while I may draw 500w at the wall, that my PC doesn't emit 500w of heat. Electricity is not energy, therefore the law of conservation of energy does not apply.
Click to expand...

So you're somehow different than the rest of the entire universe?

Please explain how your drawing 500 watts yet not turning 500 watts into heat.
Please, because there would be a Nobel Peace Prize in your inbox considering you've just permanently fixed global warming, the energy crisis and Einstein's theory of everything.
 
So - about those CrossFire drivers then. Anybody try them out on Vega 56/64? I'm curious to know how they perform.
 
griff30 said:
So you're somehow different than the rest of the entire universe?

Please explain how your drawing 500 watts yet not turning 500 watts into heat.
Please, because there would be a Nobel Peace Prize in your inbox considering you've just permanently fixed global warming, the energy crisis and Einstein's theory of everything.
Click to expand...

It's called conversion. Did you know that an 80w Incandecent light bulb doesn't dissipate 80w of heat? It dissipates less, because there has also been a conversion of electrical energy to light! I'll take my Nobel now.
 
alxlwson said:
It's called conversion. Did you know that an 80w Incandecent light bulb doesn't dissipate 80w of heat? It dissipates less, because there has also been a conversion of electrical energy to light! I'll take my Nobel now.
Click to expand...
dunce-cap.jpg


When you are done, you can read about electronics.


Back OT and with any luck, with a few who could figure out electricity is not "magic".


I don't suppose there will be a review of a crossfire system as the cost is ridiculous.
Why only 2 way CF?
 
griff30 said:
dunce-cap.jpg


When you are done, you can read about electronics.


Back OT and with any luck, with a few who could figure out electricity is not "magic".


I don't suppose there will be a review of a crossfire system as the cost is ridiculous.
Why only 2 way CF?
Click to expand...


Deal with them on a daily basis. I troubleshoot and rebuild my fair share of VFD's in the 500kw range.


I think only 2-way because it seems nobody actually wants to support it. It's a PITA for game devs and the driver teams alike. Then couple that with the fact that it never truly took off, just not worth the effort
 
alxlwson said:
You're confusing eelectricity with electrical energy. Extremely common mistake.
Click to expand...

I am, and have been throughout, discussing electrical power.

When we're talking about how much power a PC draws (that 500W number), we're referring to electrical power. If we'd like to use energy terms instead of power, then let's say 1 hour of operation, 500 watt-hours.
That amount of electrical energy does have to follow the laws of conservation of energy, because as the name indicates, it's energy.

Thus again, when 500W of electrical power comes into the closed system, it doesn't vanish, and it still exists in some form. And as it turns out, it's really 500W of heat. There is a small fraction emitted as non-heat radiation, but as it turns out, even that ends up as heat in short order. Light for example - what happens to it? Does it just vanish? No, it strikes some surface and get absorbed - and of course then it is heat.
 
PhaseNoise said:
I am, and have been throughout, discussing electrical power.

When we're talking about how much power a PC draws (that 500W number), we're referring to electrical power. If we'd like to use energy terms instead of power, then let's say 1 hour of operation, 500 watt-hours.
That amount of electrical energy does have to follow the laws of conservation of energy, because as the name indicates, it's energy.

Thus again, when 500W of electrical power comes into the closed system, it doesn't vanish, and it still exists in some form. And as it turns out, it's really 500W of heat. There is a small fraction emitted as non-heat radiation, but as it turns out, even that ends up as heat in short order. Light for example - what happens to it? Does it just vanish? No, it strikes some surface and get absorbed - and of course then it is heat.
Click to expand...

Agreed.

I will also say this -- I assumed that computers had some form of efficiency, and I am incorrect in that assumption. They really aren't efficient at all. Basically all Pin = Pout as thermal energy. I was thinking "work" being, lets just say 80% efficient, and then the rest loss in hysteresis. So I was assuming a 20% loss due to hysteresis of Pnom. I never really considered how inefficient that computers are, and that there really is no hysteresis, but complete inefficiency (thinking power factor). That is a lot of wasted energy. Typically Work is the result of Power minus loses. In the case of computers, and perhaps this is more an issue with digital electronics in general, there isn't any actual Work being done, hence a computer being for all purposes, a completely resistive load.
 
Last edited:
alxlwson said:
Agreed.

I will also say this -- I assumed that computers had some form of efficiency, and I am incorrect in that assumption. They really aren't efficient at all. Basically all Pin = Pout as thermal energy. I was thinking "work" being, lets just say 80% efficient, and then the rest loss in hysteresis. So I was assuming a 20% loss due to hysteresis of Pnom. I never really considered how inefficient that computers are, and that there really is no hysteresis, but complete inefficiency (thinking power factor). That is a lot of wasted energy. Typically Work is the result of Power minus loses. In the case of computers, and perhaps this is more an issue with digital electronics in general, there isn't any actual Work being done, hence a computer being for all purposes, a completely resistive load.
Click to expand...

Chips are either horribly inefficient wires, or really efficient electricity to heat converters. :)

That's basically the gist of it - there's no actual "work" being done from a physics definition. The switches all being in one state is the same as being in another, in terms of system energy level. Thus even with switches being in different states, there's no inherent energy needed.
But things aren't ideal, and there's a little loss to heat to push charge around to get all those switches in the next correct state due to internal resistances. And the dang things all leak, at least a little. So you get nickle-and-dimed times a few billion, heh.

Anyway, sorry everyone for this birdwalk. Hopefully it was interesting to someone.
 
PhaseNoise said:
What's a few hundred watts difference, I always say.
Click to expand...


lol i'd take it, electricity is 1/4 the price of natural gas here so it would be worth having the space heater for me. it costs me about ~140 dollars a month just to run my furnace.. but then again a space heater is 30 dollars and still puts out the same heat whether i'm using my computer or not so...
 
sirmonkey1985 said:
lol i'd take it, electricity is 1/4 the price of natural gas here so it would be worth having the space heater for me. it costs me about ~140 dollars a month just to run my furnace.. but then again a space heater is 30 dollars and still puts out the same heat whether i'm using my computer or not so...
Click to expand...

At a recent cocktail party I brought up how we just finished having a good sized solar array installed, and it was great to not have soul-eating utility bills anymore. I lamented the winter still hits us a bit since solar is just electricity. For fun, being the wildly sarcastic ass I am, said with a straight face "At least until we get our solar natural gas installed next year".

Now, this is where you'd hope for a chuckle, or at least grimace at a bad joke. Sadly, the only person who got it right away was my wife (100x smarter than me) who shot me a rather dark look from across the house, but remained silent out of sheer amusement. One person said "nice!", and one other said "oh so they finally got that working?"

Now I like these people. I know not everyone is into the same things, but a little piece of me died that day, and I imagined a tiny tear in a science teacher's eye, somewhere.

Then I got to thinking - if I was an Enron-esque rat bastard, how many people could I get signed up before I inevitably fled to Brazil with a lot of money? Good thing I'm a mild engineer. I guess.
 
You must log in or register to reply here.
Back
Top