AMD Ryzen 9 7950X prototype demolishes the 5950X in Cinebench R23

emmzo

Posts: 806   +1,240
Nah, you can turn your thermostat down in the winter and save some money. lol. But seriously, I don't like the state of energy usage. I think I would settle for new generation chips that perform at the same levels but half the energy usage. That would be AWESOME.
Yep, I would love that too and it used to be like that, to some extent. Idk what changed. Now you get 20-30% more juice and double consumption? Wtf? GPUs, the same. It's not like "oh, 300W(cpu+gpu) more, I'm going bankrupt". But isn't efficiency better?
 

takaozo

Posts: 501   +802
The real problem is not if you go bankrupt or not for a few watts more, but the cooling and power required. I'm not in the mood to change case+psu+cooling just for a new gpu or cpu. I got a decent cpu that dont require exotic cooling and same for gpu. For the gpu I did a replacement, 92x25 fans and custom shroud.
Now all is dead silent and no matter what I do it wont get past 65-70 celsius,
 

mattferg

Posts: 236   +239
Every leaked specification I've seen so far indicates a minimum TDP of 105w; the 65w tier has been eliminated.

If you consider the EV in isolation, sure. But when you realize the electricity for that EV is in most cases generated from a thermal power plant with generation, transmission, and distribution losses, plus charging losses from wall socket to EV, plus the amortized additional energy costs of manufacturing the vehicle, the equation is much more nearly balanced. EVs usually win, but by a much smaller margin, and in some countries, they wind up generating more emissions overall, not less.

Sorry, but you're completely, utterly wrong on this one. Transmission and distribution losses are miniscule compared with transporting petrol, for a start.

But on to the actual numbers - if you compare input kilojoules, an EV with power generated from a fossil fuel plant is around 45% efficient, with 45% of the drop coming from the generation and just 10% from transmission.

To put this in perspective - the current Formula 1 engines just about hit 50% efficiency, and they're some of the most efficient and complicated in the world - no road car has an MGU-H and even MGU-Ks are uncommon. Regular cars are between 10-20%. It isn't even close between EVs and gas cars, it's not REMOTELY balanced.

And as for "more emissions in some cases" this is woefully wrong as well. Crucially, due to the size and portability requirements, ICE cars are far less efficient at reducing emissions than a power station the size of a village is. Gas power plants are anywhere from 3-4x more efficient at both generating power AND reducing emissions vs a car engine - it helps when you can literally have a filtration system the size of a building. And, crucially, they're generating emissions far away from your home, not right outside of the front door.

This is before we even factor in the fact that a significant portion of the grid doesn't come from fossil fuels, which absolutely skyrockets both the efficiency and emissions stats of an EV vs an ICE car, given some input power will be from both renewables and nuclear.

So no, it's not even close when you compare EVs and ICE cars, but I should've known you were somewhat biased and clueless from the "greenies" comment.
 

Endymio

Posts: 1,918   +1,997
Sorry, but you're completely, utterly wrong on this one.
You managed to get every single fact in that lengthy post incorrect. Let's take them one by one, shall we?

Transmission and distribution losses are miniscule compared with transporting petrol...
Electrical transmission/distribution losses vary by state and country. In the US, it averages around 7%, but some states (like ID) break 13% losses, whereas some countries (like Haiti and Libya) actually lose more than 50% of all electricity generated. This is on top of the 10 to 22% loss from charging the EV itself.


Whereas distribution of gasoline involves only a 0.5% to 1.5% loss, depending on tanker truck routes, while "charging" (refueling) has no appreciable losses whatsoever.

Regular cars are between 10-20% [efficient]. It isn't even close between EVs and gas cars, it's not REMOTELY balanced.
An engine from the 1920s might have averaged 10% efficiency. Modern engines, however, do much better:

"...45 percent is a new record for passenger car engines. By way of comparison, petrol engines typically have efficiencies of 35 to 40 percent.."


crucially, [power plants] are generating emissions far away from your home, not right outside of the front door.
We were speaking of carbon emissions, which purports itself as a global issue. You realize that CO2 emitted in China is no different than CO2 emitted "at your front door", right?

So no, [emissions] are not even close when you compare EVs and ICE cars....
Wrong yet again. As I stated earlier, it varies dramatically by location. This study for instance, calculated the emissions for an EV charged in Poland at 180 g/km, substantially higher than the 163 g/km for a ICE vehicle:


And, as I said, EVs do better in other areas, and usually do outperform ICEs ... but the margin is significantly smaller than you believe.

but I should've known you were somewhat biased and clueless from the "greenies" comment.
In light of your "clueless" remarks above, I can only respond "educate yourself".
 

mattferg

Posts: 236   +239
You managed to get every single fact in that lengthy post incorrect. Let's take them one by one, shall we?

Electrical transmission/distribution losses vary by state and country. In the US, it averages around 7%, but some states (like ID) break 13% losses, whereas some countries (like Haiti and Libya) actually lose more than 50% of all electricity generated. This is on top of the 10 to 22% loss from charging the EV itself.


Whereas distribution of gasoline involves only a 0.5% to 1.5% loss, depending on tanker truck routes, while "charging" (refueling) has no appreciable losses whatsoever.

An engine from the 1920s might have averaged 10% efficiency. Modern engines, however, do much better:

"...45 percent is a new record for passenger car engines. By way of comparison, petrol engines typically have efficiencies of 35 to 40 percent.."


We were speaking of carbon emissions, which purports itself as a global issue. You realize that CO2 emitted in China is no different than CO2 emitted "at your front door", right?

Wrong yet again. As I stated earlier, it varies dramatically by location. This study for instance, calculated the emissions for an EV charged in Poland at 180 g/km, substantially higher than the 163 g/km for a ICE vehicle:


And, as I said, EVs do better in other areas, and usually do outperform ICEs ... but the margin is significantly smaller than you believe.

In light of your "clueless" remarks above, I can only respond "educate yourself".

Wow, you straight up just made up facts to defend yourself 🤦🏻‍♂️ EVs don’t lose 20% charging, this is laughable. As for “every fact wrong” I quoted 5% transmission loss and you literally came back with 7. Peak cringe.

1.5% loss transmitting gas? Riiiight. Because going from an oil well, refinery, to each individual petrol station only loses 1.5%? Hahahahaha.

We weren’t talking about carbon emissions - we were talking about emissions. FFP2, NO, sulfur, all relevant. But please narrow and rewrite the discussion to hide your blatant errors.

The study you referenced re emissions has been disproven at peer review as biased.

So again, you managed to prove nothing I said was wrong, despite a post which clearly went to a lot of effort, and confirmed several things I said were correct:

Next time, actually go educate yourself before embarrassing yourself further. I don’t need to reply any more - you did all my work for me. Awkward.
 

Endymio

Posts: 1,918   +1,997
Wow, you straight up just made up facts to defend yourself 🤦🏻‍♂️ EVs don’t lose 20% charging, this is laughable.
I stated they can lose from 10-22%, which is correct. This site, for instance, measured 12% charging loss on a (more efficient) fast-DC charge:


Here is Car and Driver, confirming the same:

" Generally speaking, your EV may use 12 to 15 percent more energy than what you add to your battery. That number could be lower or higher depending on charging conditions."


Here's a scientific study that measured a much larger loss:

"When charging or discharging electric vehicles, power losses occur in the vehicle and the building systems supplying the vehicle. ... Under the conditions studied, measured total one-way losses vary from 12% to 36%,


And ThinkEV agrees as well. Since your math is a little shaky, 84% efficiency translates to a 16% loss:

"The average energy-efficiency of Level 1 charging, according to scientific calculations, is roughly 84%..."


Yes, you got your facts wrong. Admit it, and move on.
 

Darth Shiv

Posts: 2,324   +851
Daily reminder that if the difference between a 100 watt and 200 watt cpu makes you worry that much about energy usage, you better never look at the power used to drive air conditioning or those new fangled electric cars greenies just love.
EVs are massively more efficient than combustion engines especially when coupled with renewables sourced charging. You need to factor in the extraction, refining waste and cost and losses from each step of their fuel costs plus of course the vehicle construction. At least 50% better over the life of the vehicle. A bit of perspective needed thanks.

Add on top of that the massive reduction in chemical emissions from OPEX not burning fuels.

Ps the fossil fuel industry has been hiding toxic byproduct data for decades. It is far higher than you think it is. How do I know? I've worked with sampling labs on the data.
 

Hodor

Posts: 394   +283
Next gen being 20-30% faster? Who would have thought? But kidding aside, what's with the huge power usage? CPUs and GPUs look hungrier than ever and just in time for the upcoming energy crisis.

That's because computer will be the only affordable source of heating this winter. 240W for an entire apartment is not much, but still better than the previous generation,.
 

mattsie

Posts: 68   +40
This was already explained, the CPU won't run at 170W even. It was said that the CPU has some quirks they need to polish out.

Somebody did the same test with the BIOS fix and fixed the problem and the CPU ran at 60-65c afterwards with a water cooler.

Also, these are all unconfirmed tests and should be discarded until the release and proper tests are done.