The 7950X it s ahead of a 13900K, and even of a 13900KS using an AIO...
THG do some artificial average where they use CB, Y-Cruncher and Povray in single thread, all softwares that are never used in ST, beside Povray doesnt use AVX2 with whatever Zen, wich cut its perfs by 15% in this test.
Intel Core i9-13900KS mit 6,0 GHz im Test
Der Core i9-13900KS ist der erste Intel-Prozessor mit bis zu 6,0 GHz ab Werk. Die dritte Special Edition in Anwendungen und Spielen im Test.
Computerbase also use Povray with AVX2 non functional for Zen, beside they use CB R20 as well as R23, wich greatly advantage Intel in the average, but that s not enough to hedge out the 7950X.
The only area where Intel win massively is in inefficency...
Default behavior for a 7950X if you install it into any X670E or B650E motherboard is to run 230W and 95 C. They'll both technically set things alight.
That's a bold claim going by a single source that shows it winning by 1% in multithread performance while losing by 10% in single thread performance. There's other sources that compare the processors such as TPU. For multithreading performance it's close enough in raw performance that it depends on the workload and overall they're effectively tied.
14900K is going to hold the performance crown for the next 6 months at least.
Intel manage to keep up thanks to vaster amounts of power.
Their power management is such that the 13900K FI manage to stick to 253W whatever the MT load.
AMD s on the other hand has a losely implemented algorithm, FI the 7950X use 205W in Cinebench and no more than 189W in Handbrake, all softs where RPL is always at full tilt power wise.
At the end it s an out of the box overclocked to the max SKU, and next Intel CPUs dont seem to moderate this orgy of power comsumption.
Where is the Intel cross statement that's supposed to follow in a Intel Thread? This is borderline trolling in an Intel thread, even if its true.
Yeah, till Zen5 comes out.
Saylick
Diamond Member
- Sep 10, 2012
- 4,052
- 9,472
- 136
Trolling in an intel thread is not allowed.
No, it doesn’t. It’ll run to 230W assuming it’s not thermally limited. Multiple reviewers show it pushing 230W in cinebench and blender. Furthermore, I owned a 7950X and with my EK 360 AIO it’d hit 230W in benchmarks that had 100% utilization across all 16 cores. The 7950X PPT is 230W at 160A.
No one cares this is not AMD thread, and no 7950x do not work on 140 wats, and even if you set it, you can do same to 14900k
That s wrong, it s temp limited to 95°C whatever the power in use, if it reach 95°C it will limit the power even if the full TDP is not exhausted, take a look at Computerbase review, they measure 205W in CB and they stated that they couldnt overclock the chip unless they unervolt it to reduce power and hence temperature.
FTR they measured 205W in CB, 189W in Handbrake H264 and 195W in Prime 95.
On the other hand RPL will gladly reach and even surpass 100 °C, that s why it can be brought to 253W.
Beside if it was to use 230W in CB FI then a full system would consume just 30W more than a full system 13900K, yet 7950X full system use 320W while the 13900K system is at 394W, that s 74W difference at the main, and accounting for the PSU + VRM losses that make a 55-60W difference at the CPU level, so since the 13900K is at 253W that put the 7950X at 200W.
The numbers are all on the link i posted, you re just relying on hear say rather than on professionaly made measurement, here it is again, look at the graph "leistungsaufname", there s CB and Prime 95 for power comsumption.
Intel Core i9-13900KS mit 6,0 GHz im Test
Der Core i9-13900KS ist der erste Intel-Prozessor mit bis zu 6,0 GHz ab Werk. Die dritte Special Edition in Anwendungen und Spielen im Test.
I’m not sure what you mean. What I described is the typical behavior for majority of people who buy a 7950X.
I’m going by TPU, GamersNexus, Hardware Unboxed, Tom’s Hardware and I’m sure others (that’s all I bothered checking). In the case of GamersNexus they measured power consumption at the EPS and got 250W and 295W in a blender workload for the 7950X and 13900K respectively. You’re quoting me a single source that probably didn’t have adequate hardware for cooling.That s wrong, it s temp limited to 95°C whatever the power in use, if it reach 95°C it will limit the power even if the full TDP is not exhausted, take a look at Computerbase review, they measure 205W in CB and they stated that they couldnt overclock the chip unless they unervolt it to reduce power and hence temperature.
FTR they measured 205W in CB, 189W in Handbrake H264 and 195W in Prime 95. Beside if it was to use 230W in CB FI then a full system would consume just 30W more than a full system 13900K, yet 7950X full system use 320W while the 13900K system is at 394W, that s 74W difference at the main, and accounting for the PSU + VRM losses that make a 55-60W difference at the CPU level, so since the 13900K is at 253W that put the 7950X at 200W.
The numbers are all on the link i posted, you re just relying on hear say rather than on professionaly made measurement, here it is again, look at the graph "leistungsaufname", there s CB and Prime 95 for power comsumption.
Intel Core i9-13900KS mit 6,0 GHz im Test
Der Core i9-13900KS ist der erste Intel-Prozessor mit bis zu 6,0 GHz ab Werk. Die dritte Special Edition in Anwendungen und Spielen im Test.www.computerbase.de
Tjmax is 100 C, it will throttle to prevent exceeding that temperature. The 13900K also has less heat density and a more efficient IHS so it’s capable of running more power so it’s not apples to apples.
Just one thing to begin with, TPU, GN and other HWUB are amateurish compared to Computerbase, they do not have 1/10 of their measurement gear, Computerbase has an automated benchmarking tool, no one provide such exhaustive measurements, they are the only ones who test at different power levels and with different softwares.
Look at the power curves they provide in the 13900KS review, FI in Handbrake they show the 13900K starting at 300W CPU power and finishing the bench at 265W, the 13900KS start at 315W and finish at 275W, and that s at stock settings
You can imagine to what extent intel has gone to match AMD, their CPU exceed the 253W power limit even at stock, not counting that it s litteraly a fraud to brand them as 125W since they are always way above this power.
On the same test the 7950X use 189W and still outperform both.
Think about it, if the 7950X was using 230W to match a 253W CPU then it would mean that it has only 10% better perf/watt, wich is the myth that people like you believe, numbers say otherwise when they are measurement made by competent people, the sites you cited dont know jack, there s one of them that use peak power as "measurement", wich is total non sense.
I put it back under your eyes, tell me if there s a single site that provide as much curves and measurements than this one, you even have the frequency in the course of the bench.
Intel Core i9-13900KS mit 6,0 GHz im Test
Der Core i9-13900KS ist der erste Intel-Prozessor mit bis zu 6,0 GHz ab Werk. Die dritte Special Edition in Anwendungen und Spielen im Test.
Power efficiency isn’t a single number, it’s a curve. Your own source shows the 7950X having a 12% performance advantage when both are limited to 142W. That’s assuming those numbers are accurate and include the 7950X’s SOC power (they usually don’t). So yeah? Roughly 10% better perf/watt.Think about it, if the 7950X was using 230W to match a 253W CPU then it would mean that it has only 10% better perf/watt, wich is the myth that people like you believe, numbers say otherwise when they are measurement made by competent people, the sites you cited dont know jack, there s one of them that use peak power as "measurement", wich is total non sense.
Intel Core i9-13900KS mit 6,0 GHz im Test
Der Core i9-13900KS ist der erste Intel-Prozessor mit bis zu 6,0 GHz ab Werk. Die dritte Special Edition in Anwendungen und Spielen im Test.
- Sep 28, 2005
- 21,067
- 3,574
- 126
If i see one more AMD post from anyone following this post without a valid reason, i will issue out more infractions.
If you want to talk about 7950X goto the Zen4 thread unless you have a valid reason in the post for talking about it.
If you want to cross compare Zen5, well its not even out yet, so why are you bringing in Zen5 and then cross comparing it to RPL and SPL? Take it to Zen 5.
I don't want to see Zen 5 when the CPU is not even out in an Intel Thread.
Otherwise keep on track.
And if i see one post with just AMD without any valid cross comparison to a Intel processor, you can also expect a infraction REALLY FAST.
And yes same goes with AMD threads, don't spam RPL or SPL in those threads without a valid reason for comparison, otherwise you'll see me in there too.
Moderator Aigo
If you want to talk about 7950X goto the Zen4 thread unless you have a valid reason in the post for talking about it.
If you want to cross compare Zen5, well its not even out yet, so why are you bringing in Zen5 and then cross comparing it to RPL and SPL? Take it to Zen 5.
I don't want to see Zen 5 when the CPU is not even out in an Intel Thread.
Otherwise keep on track.
And if i see one post with just AMD without any valid cross comparison to a Intel processor, you can also expect a infraction REALLY FAST.
And yes same goes with AMD threads, don't spam RPL or SPL in those threads without a valid reason for comparison, otherwise you'll see me in there too.
Moderator Aigo
Last edited:
NVIDIA Grace Hopper Superchip Sweeps MLPerf Inference Benchmarks
NVIDIA GH200, H100 and L4 GPUs and Jetson Orin modules show exceptional performance running AI in production from the cloud to the network’s edge.
blogs.nvidia.com
Intel Shows Strong AI Inference Performance
www.intel.com
NVIDIA shows off H100 absolute performance.
Also Intel does impressive Gaudi 2 perf/cost.
However, GH200 doesn't look so good.
I estimate Sapphire Rapids + H100 server system keeping at 96.5%~ performance of GH200 system.
In addition, SPR + H100 server is probabry a relatively inexpensive system in TCO.
Because H100's crazy price-tag, I suspect that GH200 is a failure product and MI300 cloud lose itself position between Gaudi 2 and H100 as AI chip.
Not at all, read that to match a 142W 7950X, that is, at isoperf, the 13900K require 200W or so, so that s 40% better perf watt at isoperf.
That being said ARL is supposed to score 15-20% better than a 13900K at 250W, that is , at iso power, at isoperf it should require something like 200W, so it should more or less match a 7950X perf/watt at isoperf but the contender in 2024 will not be Zen 4.
Well, it looks like 2023 is the first time in many years Intel is going to have a comfortable lead both in desktop cpu performance and laptop power-efficiency. Next 6 months is gonna be very interesting for Intel...
8 more days to go...
8 more days to go...
That’s literally what it shows. You can manipulate it and twist it by selecting certain points on the efficiency curve that suit your argument but when both are power limited to 142W it shows a 10% performance advantage over the 13900KS and an 11% performance advantage over the 13900K.
To prove the point, your data shows the 13700K at 142W outperforming the 12900KS at 241W! Same core counts and everything. I guess RPL improved efficiency by 70% in a single generation!
Alternatively, at iso-power (125W) the 13700K has 7% more performance than the 12900K.
Which one of those seems like a more realistic assessment?
Last edited:
You can either compare them at iso-power or iso-perf.
I personally think iso-power is a better comparison, because achieving a few more % of performance can significantly affect the efficiency in a negative way.
I personally think iso-power is a better comparison, because achieving a few more % of performance can significantly affect the efficiency in a negative way.
Lunar lake is the big one from insider sources will beat apple in all stacks... beating apple in power efficiency is the goal for intel.. apple are undisputed in performance per watt on laptop
But that s not how things work.
When perf increase by 10% perf/watt decrease by at least 10%, so that s not a good comparison since the better performing chip is at a disadvantage in such a comparison, best is to compare at equal perf to see what is the real efficency difference.
That s surely due to the fact that there s surely some short turbo at work even when the TDP is fixed at 142W, we can see that the 13700K@253W perform only 9% better than when set at 142W, wich is not logical, the difference should be more.
As said there s surely some turboing at work, at 125W the difference between the 13700K and 12900K should be bigger than 7%, more like 10% at least.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 22K
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 384
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
Facebook
Twitter