Only if you have 120v to play with. Real countries run 13A on a 240v supply giving 3,120W per outlet.
Being in Europe has a few benefits in that department.
You want to make absolutely sure your installation can handle that load long term. We've had "issues" here with people plugging in their emergency charger for their electric car into regular outlets (usually in summerhouses with no chargers for miles and miles around), with fiery results.
Cables were found to hit 108C after 15 minutes...
That just sounds like faulty (or old) wiring. I had to use the granny charger for my EV for a few weeks until the 7.7KW charger was installed and had no issues with it. The cable from my fuse board to that fast charger is one chunky mofo though.
Power consumption isn't a simple statistic. Yet, reviewers and readers usually tend to try to use just one single number for a complex discussion. CPU power consumption depends on workload, cooling system, power supply, time since the start of heavy work, room temperature, etc. All those things can't be boiled down to something like "125 W". A much more accurate description might be:Indeed. So what's with the FUD about power consumption.
Power consumption isn't a simple statistic. Yet, reviewers and readers usually tend to try to use just one single number for a complex discussion. CPU power consumption depends on workload, cooling system, power supply, time since the start of heavy work, room temperature, etc. All those things can't be boiled down to something like "125 W". A much more accurate description might be:
"Up to 300 W for up to 56 seconds, in a AVX-512 workload, and assuming the default heat sink was removed and replaced with a much more powerful cooling system, and the CPU was put into a performance oriented motherboard, otherwise up to 125W."
You've seen reviews with runs with both chips at 125 W with the default cooler both side-by-side? If so, please link. If not, then you don't have any reason to say it is indisputable.Your statement actually sounds like more an attempt to obfuscate perf/W.... which is exactly what "TheElf" tried to do by claiming you cannot conclude Rocketlake takes more power to hit the same performance as Cometlake, which at this point is indisputable.
You've seen reviews with runs with both chips at 125 W with the default cooler both side-by-side? If so, please link. If not, then you don't have any reason to say it is indisputable.
Just look at the Agisoft result vs the power consumption and draw your own conclusions. Or do the math on per core power versus Cometlake... the answer is obvious.
Ian did Intel a huge favor by not doing a power-tied comparison with Rocketlake.
I think reviewers should always do both. (1) Default settings, default cooling if supplied or cooler rated at TDP if not, in default box. This is what the vast majority of users will experience. (2) Then all-out best cooler, best motherboard, what can the CPU possibly do in the best light. This is great for enthusiasts that want that last drop of performance. Performance would be worse in #1 than #2. But power would be much lower too. You seem to want to bash a CPU for low performance in #1 but high power usage in #2: a situation that likely never occurs outside a few reviews.Just look at the Agisoft result vs the power consumption and draw your own conclusions. Or do the math on per core power versus Cometlake... the answer is obvious.
Ian did Intel a huge favor by not doing a power-tied performance comparison. Would have been ugly even against Cometlake, never mind Zen 3/M1.
I think reviewers should always do both. (1) Default settings, default cooling, in default box. This is what the vast majority of users will experience. (2) Then all-out best cooler, best motherboard, what can the CPU possibly do in the best light. This is great for enthusiasts that want that last drop of performance. Performance would be worse in #1 than #2. But power would be much lower too. You seem to want to bash a CPU for low performance in #1 but high power usage in #2: a situation that likely never occurs outside a few reviews.
I think reviewers should always do both. (1) Default settings, default cooling, in default box. This is what the vast majority of users will experience. (2) Then all-out best cooler, best motherboard, what can the CPU possibly do in the best light. This is great for enthusiasts that want that last drop of performance. Performance would be worse in #1 than #2. But power would be much lower too. You seem to want to bash a CPU for low performance in #1 but high power usage in #2: a situation that likely never occurs outside a few reviews.
I didn't assume. In fact I made no statement at all regarding that. I asked you for links to see what it is for myself. You are the one just stating it.Heh... why do you assume RL can beat CL within the power envelope of #1?
As for #2, I don't even bother discussing AVX-512 power, I have openly stated that AVX-512 is a joke so it would be rather unfair to hold power consumption for a joke instruction set against anyone. You don't need AVX-512 to see that RKL perf/W is still in the dumpster.
I didn't assume. In fact I made no statement at all regarding that. I asked you for links to see what it is for myself. You are the one just stating it.
So, you are not including AVX-512? Lets see why that might be from Anandtech's preview.
That seems to me like you exclude the cases where Rocket Lake is way more power efficient then conclude that it therefore must be less power efficient.
- Power: https://images.anandtech.com/graphs/graph16535/121878.png
- AVX Performance: https://images.anandtech.com/graphs/graph16535/121883.png
- 10700k: 5211 MMovements/sec with 204.79 W power consumption. 25.4 MMovement / joule
- 11700k: 32845 MMovements/sec with 291.68 W power consumption. 112.6 MMovement / joule
Good catch. The 10700 does come with a cooler, the 10700k does not. I should clarify my post.Do the desktop i7/i9 CPUs come with a cooler?
Sure, Intel is way more efficient in doing something that nobody cares about. I will definitely give you that, that is Intel's specialty.
If I am going to refrain from using a joke instruction set to criticize power consumption, at least you can refrain from using that same joke instruction set to brag about relative efficiency: it is like bragging about being the best at digging a ditch in the middle of nowhere.
Quick tell AMD to take AVX-512 out of Zen 4, since you clearly know so much more than AMD!Sure, Intel is way more efficient in doing something that nobody cares about. I will definitely give you that, that is Intel's specialty.
If I am going to refrain from using a joke instruction set to criticize power consumption, at least you can refrain from using that same joke instruction set to brag about relative efficiency: it is like bragging about being the best at digging a ditch in the middle of nowhere.
The difference is AMD is planning to launch it in the future when there *might* be more use cases and also they plan to do it on appropriate process node so that the whole chip is not burning a hole in space-time continuum. It is sad how behind AMD they are currently, in both process node and core development.Quick tell AMD to take AVX-512 out of Zen 4, since you clearly know so much more than AMD!
It couldn't possibly be used in things like photoshop, video editing, compression, science, etc.
AVX-512 being able to deliver improvements in performance still doesn't improve it being a mess of an instruction design. As I stated years ago I still wish AMD would support something more akin to SVE instead or alongside AVX-512.Quick tell AMD to take AVX-512 out of Zen 4, since you clearly know so much more than AMD!
It couldn't possibly be used in things like photoshop, video editing, compression, science, etc.
Quick tell AMD to take AVX-512 out of Zen 4, since you clearly know so much more than AMD!
It couldn't possibly be used in things like photoshop, video editing, compression, science, etc.
The silicon area ROI is not not even remotely justifiable considering those workloads are far better sent to a compute accelerator. AMD could have resisted this ISA conformity... but I don't run AMD, so whatever.
You are correct that it is a mess. Removing features in later releases makes it a terrible rollout.AVX-512 being able to deliver improvements in performance still doesn't improve it being a mess of an instruction design. As I stated years ago I still wish AMD would support something more akin to SVE instead or alongside AVX-512.
Correct me if I'm wrong, but isn't AVX-512 pretty tiny in silicon area? I thought I saw a die shot with it called out and it wasn't much of the chip, but I could be wrong there. You are correct that AVX-512 is better as a separate accelerator. That is Intel's long term plan to mix-and-match chiplets with foveros on a CPU. But they have to start somewhere to get software started using it.The silicon area ROI is not not even remotely justifiable considering those workloads are far better sent to a compute accelerator. AMD could have resisted this ISA conformity... but I don't run AMD, so whatever.
There's a bunch of HPC customers that like AVX-512. Is it that too niche of a market to add it just for them?