AMD's Future Chips & SoC's: News, Info & Rumours.

goldstone77 · Dec 27, 2017

juanrga :

Not all notebooks with the AMD Ryzen™ mobile APU will offer the necessary thermal solution to enable the performance upside of mXFR, but the HP ENVY x360 featuring the AMD Ryzen™ Processor with Radeon™ Vega Graphics is the first solution to do so. Users will look for “amplified mXFR performance” in the marketplace should they desire a laptop that offers this capability.

So, 25W TDP?, and what's Intel's excuse? You are lying to yourself if you think Intel only pulls 15W of power, because they market it at a 15W TDP.

https://www.laptopmag.com/reviews/laptops/hp-envy-x360-15t
Ryan Shrout loves his HP 360X and recommends it as his pick of the week!
Picks of the Week:
1:19:10 Ryan: HP Envy x360 Ryzen 5 If that Intel shill likes it, than it has to be amazing!
https://youtu.be/ou2M5NKO25A?t=4752

Edit: You have clearly never used the piece of garbage HSF that intel sends with it's processors that won't keep their CPU's cool at their rated TDP!

goldstone77 · Dec 27, 2017

According to AMD, "TDP is essentially the maximum sustained power a processor can draw with “real world” software while operating under defined temperature and voltage limits."

According to Intel, "TDP represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload."

Edit: When it comes to AMD or Intel TDP is just a marketing term! They define their own parameters that constitutes what their TDP will be.

juanrga · Dec 27, 2017

goldstone77 :

Point 7) above. It is not Intel vs AMD issue, but AMD vs AMD.

juanrga · Dec 27, 2017

goldstone77 :

The Techreport review isolated the different battery and screen to compare just the SoCs. The laptops compared by laptop aren't identical.

The Intel version of the Envy x360 has 16GB RAM. The AMD version has only 8GB. More RAM implies more power consumption. The Intel version has a much more bright screen:

The Intel Envy x360 emits a paltry 186 nits of brightness, while its AMD counterpart maxes out at a shockingly low 128 nits.

The more bright is the screen the more power it consumes. And the Intel-based laptop had a screen is 45% more bright.

So despite having higher power consumption from more RAM and brighter screen the Intel-based laptop lasted more time, 5:49 versus 5:11 or what is the same the Intel-based laptop lasted 12% more. With same screen and same memory, the battery-life difference would probably increase to 30% or so.

goldstone77 :

Are you aware that AMD ships its '95W' CPUs with a 140W cooler and its 65W CPUs with a 95W cooler? 140W are needed for the 1800X/1700X because the real TDP is 128W and 95W cooler is needed for the 1700 because the real TDP is 90W.

adamsleath · Dec 27, 2017

juanrga :

lol

incorrect

"1700 because the real TDP is 90W" ----> incorrect.
but scour the internet for some evidence then report back with some more trolling.

also irrelevant at this point as there is no agreement here on precise definition of TDP, or testing parameters.

...and yes good luck with a stock heatsink on a coffee lake. even a non k chip is more likely to throttle under load due to overheating. 85C + ? :lol:

"This is not Intel vs AMD" - then why do you persist in trolling the AMD thread? with cherry picked and often dubious and inaccurate statements. with a specific and obvious intel bias?

it has devolved into nitpicking. it will always be intel vs amd because its the 'same' architecture x86/x64 same instruction types. with piddlingly pathetic minor differences...which are blown out of proportion here to differentiate the products to attempt to snag sales.

if it wasnt for the fact they are in the same market then there would be no banter.

ultimately customers like choice. they dont like monopolies that inevitably lead to higher prices.

both intel and amd are overpriced garbage anyway, doing similar marketing strategies and product heirarchies :lol: and attempting to diss each other at any opportunity (customer trolling / whinging)

it has been good in recent years though for customers, versus manufacturers who need built in redundancy to shift stock, as many customers are still happy with 1,2,3,4,5 and 6 year old processors.

due to your WALL. progress is so slow as to obviate the need for upgrading. - enough negativity for ya?

jdwii · Dec 27, 2017

YoAndy :

You are right 100% that is what TDP means heat dispersion NOT power consumption.Intel+Amd state this

https://www.youtube.com/watch?v=yDWO177BjZY

Here is a old Techquickie "TDP is not directly measure of power consumption" Linus Sebastian

https://www.lifewire.com/thermal-design-power-831763

"TDP stands for Thermal Design Power. And while many computer users may think it equates to the maximum amount of power a component can run at, that isn't the case. TDP is technically the max amount of power the cooling system needs to dissipate in order to keep the chip at or below its maximum temperature."

https://www.windowscentral.com/what-tdp-and-why-should-you-care-about-it

"TDP doesn't equate to how much power will be drawn by the component in question, but that doesn't mean you can't use the value provided as an estimation."

goldstone77 · Dec 27, 2017

juanrga :

So, I need to buy at least an AMD cooler to try and keep an Intel CPU cool, because the HSF they ship with their CPU's don't work at the rated TDP!

goldstone77 · Dec 27, 2017

juanrga :

Ryan Shrout loves his HP 360X and recommends it as his pick of the week! So, I guess he doesn't care about the battery life using a monitor, because he can actually game with the AMD unit in games where the Intel unit with it's igpu is unplayable.
Picks of the Week:
1:19:10 Ryan: HP Envy x360 Ryzen 5
https://youtu.be/ou2M5NKO25A?t=4752

adamsleath · Dec 27, 2017

jdwii :

YoAndy :

You are right 100% that is what TDP means heat dispersion NOT power consumption.Intel+Amd state this

https://www.youtube.com/watch?v=yDWO177BjZY

Here is a old Techquickie "TDP is not directly measure of power consumption" Linus Sebastian

https://www.lifewire.com/thermal-design-power-831763

"TDP stands for Thermal Design Power. And while many computer users may think it equates to the maximum amount of power a component can run at, that isn't the case. TDP is technically the max amount of power the cooling system needs to dissipate in order to keep the chip at or below its maximum temperature."

https://www.windowscentral.com/what-tdp-and-why-should-you-care-about-it

"TDP doesn't equate to how much power will be drawn by the component in question, but that doesn't mean you can't use the value provided as an estimation."

...technically, anything below tjmax (around 100c ? on intel chip?) is fine then...which is auto shutdown territory.

and ryzen ive read is about 95C ? shutdown. with 75-80 max temp preferrable?

goldstone77 · Dec 27, 2017

aHR0cDovL21lZGlhLmJlc3RvZm1pY3JvLmNvbS9JL1EvNjg0OTYyL29yaWdpbmFsLzA0LVBvd2VyLUNvbnN1bXB0aW9uLVRvcnR1cmUucG5n

Pushing all of the Core i9-7900X’s cores with Prime95 or LuxRender propels power consumption to incredible heights. You do get 48 percent more rendering performance in LuxRender, but at the expense of 58 percent-higher power use. This approach has the elegance of a sledgehammer. Then again, if you need speed at any cost, Core i9-7900X is top-notch.

So this is surprising. Normally Intel are relatively good at their recommended TDP numbers: the ability to remove a certain amount of heat related to power consumption is something Intel either gets bang on, or has plenty of headroom. The sole Kaby Lake-X CPU that completed this test is an example: at under 60W, it is comfortably under the 112W TDP that chip has. But for the 140W Skylake-X parts, we recorded nearly 150W power consumption. Intel announced that the socket is suitable up to 165W, so it’s clear that they are pushing the frequencies here and it is going to be telling what might happen with the higher core count silicon.

A Few Words on Power Consumption
When we tested the first wave of Skylake-X processors, one of the take away points was that Intel was starting to push the blurred line between thermal design power (TDP) and power consumption. Technically the TDP is a value, in Watts, to which a CPU cooler should be designed to cope with heat energy of that amount: a processor with a 140W TDP should be paired with a CPU cooler that can dissipate a minimum of 140W in order to avoid temperature spikes and ‘thermal runaway’. Failure to do so will cause the processor to hit thermal limits and reduce performance to compensate. Normally the TDP is, on average, also a good metric for power consumption values. A processor with a TDP of 140W should, in general, consume 140W of power (plus some efficiency losses).

In the past, particularly with mainstream processors, and even with the latest batch of mainstream processors, Intel typically rides the power consumption well under the rated TDP value. The Core i5-7600K for example has a TDP of 95W, and we measured a power consumption of ~61W, of which ~53W was from the CPU cores. So when we say that in the past Intel has been conservative with the TDP value, this is typically the sort of metric we will quote.

With the initial Skylake-X launch, things were a little different. Due to the high all-core frequencies, the new mesh topology, the advent of AVX-512, and the sheer number of cores in play, the power consumption was matching the TDP and even exceeding it in some cases. The Core i9-7900X is rated at 140W TDP, however we measured 149W, a 6.4% difference. The previous generation 10-core, the Core i7-6950X was also rated at 140W, but only draws 111W at load. Intel’s power strategy has changed with Skylake-X, particularly as we ramp up the number of cores.

Even though we didn’t perform the testing ourselves, our colleagues over at Tom’s Hardware, Paul Alcorn and Igor Wallossek, did extensive power testing on the Skylake-X processors. Along with showing that the power delivery system of the new motherboards works best with substantial heatsinks and active cooling (such as a VRM fan), they showed that with the right overclock, a user can draw over 330W without too much fuss.

So for the two processors in the review today, the same high values ring true. Almost to the point of it being alarmingly so. Both the Core i9-7980XE and the Core i9-7960X have a TDP rating of 165W, and we start with the peak headline numbers first. Our power testing implements a Prime95 stress test, with the data taken from the internal power management registers that the hardware uses to manage power delivery and frequency response. This method is not as accurate as a physical measurement, but is more universal, it removes the need to tool up every single product, and the hardware itself uses these values to make decisions about the performance response.

At full load, the total package power consumption for the Core i9-7960X is almost on the money, drawing 163W.

However the Core i9-7980XE goes above and beyond (and not necessarily in a good way). At full load, running an all-core frequency of 3.4 GHz, we recorded a total package power consumption of 190.36W. This is a 25W increase over the TDP value, or a 15.4% gain. Assuming our singular CPU is ‘representative’, I’d hazard a guess and say that the TDP value of this processor should be nearer 190W, or 205W to be on the safe side. Unfortunately, when Intel started designing the Basin Falls platform, it only was designed to be rated at 165W. This is a case of Intel pushing the margins, perhaps a little too far for some. It will be interesting to get the Xeon-W processors in for equivalent testing.

Conclusion: Intel lies about their TDP numbers, and we don't have to search the web and translate a french website to prove it. Three reputable sites including Tom's Hardware in English.

YoAndy · Dec 28, 2017

goldstone77 :

Pushing all of the Core i9-7900X’s cores with Prime95 or LuxRender propels power consumption to incredible heights. You do get 48 percent more rendering performance in LuxRender, but at the expense of 58 percent-higher power use. This approach has the elegance of a sledgehammer. Then again, if you need speed at any cost, Core i9-7900X is top-notch.

So this is surprising. Normally Intel are relatively good at their recommended TDP numbers: the ability to remove a certain amount of heat related to power consumption is something Intel either gets bang on, or has plenty of headroom. The sole Kaby Lake-X CPU that completed this test is an example: at under 60W, it is comfortably under the 112W TDP that chip has. But for the 140W Skylake-X parts, we recorded nearly 150W power consumption. Intel announced that the socket is suitable up to 165W, so it’s clear that they are pushing the frequencies here and it is going to be telling what might happen with the higher core count silicon.

A Few Words on Power Consumption
When we tested the first wave of Skylake-X processors, one of the take away points was that Intel was starting to push the blurred line between thermal design power (TDP) and power consumption. Technically the TDP is a value, in Watts, to which a CPU cooler should be designed to cope with heat energy of that amount: a processor with a 140W TDP should be paired with a CPU cooler that can dissipate a minimum of 140W in order to avoid temperature spikes and ‘thermal runaway’. Failure to do so will cause the processor to hit thermal limits and reduce performance to compensate. Normally the TDP is, on average, also a good metric for power consumption values. A processor with a TDP of 140W should, in general, consume 140W of power (plus some efficiency losses).

In the past, particularly with mainstream processors, and even with the latest batch of mainstream processors, Intel typically rides the power consumption well under the rated TDP value. The Core i5-7600K for example has a TDP of 95W, and we measured a power consumption of ~61W, of which ~53W was from the CPU cores. So when we say that in the past Intel has been conservative with the TDP value, this is typically the sort of metric we will quote.

With the initial Skylake-X launch, things were a little different. Due to the high all-core frequencies, the new mesh topology, the advent of AVX-512, and the sheer number of cores in play, the power consumption was matching the TDP and even exceeding it in some cases. The Core i9-7900X is rated at 140W TDP, however we measured 149W, a 6.4% difference. The previous generation 10-core, the Core i7-6950X was also rated at 140W, but only draws 111W at load. Intel’s power strategy has changed with Skylake-X, particularly as we ramp up the number of cores.

Even though we didn’t perform the testing ourselves, our colleagues over at Tom’s Hardware, Paul Alcorn and Igor Wallossek, did extensive power testing on the Skylake-X processors. Along with showing that the power delivery system of the new motherboards works best with substantial heatsinks and active cooling (such as a VRM fan), they showed that with the right overclock, a user can draw over 330W without too much fuss.

So for the two processors in the review today, the same high values ring true. Almost to the point of it being alarmingly so. Both the Core i9-7980XE and the Core i9-7960X have a TDP rating of 165W, and we start with the peak headline numbers first. Our power testing implements a Prime95 stress test, with the data taken from the internal power management registers that the hardware uses to manage power delivery and frequency response. This method is not as accurate as a physical measurement, but is more universal, it removes the need to tool up every single product, and the hardware itself uses these values to make decisions about the performance response.

At full load, the total package power consumption for the Core i9-7960X is almost on the money, drawing 163W.

However the Core i9-7980XE goes above and beyond (and not necessarily in a good way). At full load, running an all-core frequency of 3.4 GHz, we recorded a total package power consumption of 190.36W. This is a 25W increase over the TDP value, or a 15.4% gain. Assuming our singular CPU is ‘representative’, I’d hazard a guess and say that the TDP value of this processor should be nearer 190W, or 205W to be on the safe side. Unfortunately, when Intel started designing the Basin Falls platform, it only was designed to be rated at 165W. This is a case of Intel pushing the margins, perhaps a little too far for some. It will be interesting to get the Xeon-W processors in for equivalent testing.

Conclusion: Intel lies about their TDP numbers, and we don't have to search the web and translate a french website to prove it. Three reputable sites including Tom's Hardware in English.

Are you even listening?. POWER CONSUMPTION HAS NOTHING TO DO WITH TDP. Those people are crazy no one is lying Intel is not lying. TDP Says nothing about power efficiency of a component, they are wrong, you are wrong by listening to them. TDP is the AVERAGE maximum Energy, expressed in watts, that needs to be dissipated by the cooling system.. When Intel talks about a CPU with 190W TDP, they are talking about HEAT.. TDP Stands for THERMAL DESIGN POWER.. The CPU with 190W TDP doesn't mean that the the processor will need 190W of power from the power supply, .. Instead of showcasing what the component will require as raw input, manufacturers use TDP as a nominal value for cooling systems to be designed around. Using a CPU as an example, one with higher TDP will draw more electricity from the PSU. Like I said before TDP is not a direct measure of how much power a component will draw from the power supply. I already explained just that on my previous post.

-Fran- · Dec 28, 2017

YoAndy :

Are you even listening?. POWER CONSUMPTION HAS NOTHING TO DO WITH TDP. Those people are crazy no one is lying Intel is not lying. TDP Says nothing about power efficiency of a component, they are wrong, you are wrong by listening to them. TDP is the AVERAGE maximum Energy, expressed in watts, that needs to be dissipated by the cooling system.. When Intel talks about a CPU with 190W TDP, they are talking about HEAT.. TDP Stands for THERMAL DESIGN POWER.. The CPU with 190W TDP doesn't mean that the the processor will need 190W of power from the power supply, .. Instead of showcasing what the component will require as raw input, manufacturers use TDP as a nominal value for cooling systems to be designed around. Using a CPU as an example, one with higher TDP will draw more electricity from the PSU. Like I said before TDP is not a direct measure of how much power a component will draw from the power supply. I already explained just that on my previous post.

He was being sarcastic, me thinks.

8350rocks · Dec 28, 2017

-Fran- :

YoAndy :

He was being sarcastic, me thinks.

That makes 2 of us...

goldstone77 · Dec 28, 2017

8350rocks :

That makes 3 of us hahaha! Yo, Andy I know what Thermal Design Power means, scroll up I've been posting the definition over multiple post now. If you don't like what Anandtech and Tom's Hardware said about Intel's version of TDP vs. Power consumption I suggest you take it up with the with the authors, there is a commentary section!
The Intel Skylake-X Review: Core i9 7900X, i7 7820X and i7 7800X Tested
by Ian Cutress on June 19, 2017 9:01 AM EST

Intel Core i9-7900X Review: Meet Skylake-X
by Paul Alcorn June 19, 2017 at 6:00 AM

YoAndy · Dec 29, 2017

goldstone77 :

8350rocks :

That makes 3 of us hahaha! Yo, Andy I know what Thermal Design Power means, scroll up I've been posting the definition over multiple post now. If you don't like what Anandtech and Tom's Hardware said about Intel's version of TDP vs. Power consumption I suggest you take it up with the with the authors, there is a commentary section!
The Intel Skylake-X Review: Core i9 7900X, i7 7820X and i7 7800X Tested
by Ian Cutress on June 19, 2017 9:01 AM EST

Intel Core i9-7900X Review: Meet Skylake-X
by Paul Alcorn June 19, 2017 at 6:00 AM

You got me there 😉. And is not that I don't like what they say, is that if they are saying that TDP is power consumption, what they say is wrong.

goldstone77 · Dec 29, 2017

https://www.ces.tech/
JAN 9-12, 2018
LAS VEGAS, NV

adamsleath · Dec 29, 2017

yeah we wait.

juanrga · Dec 30, 2017

adamsleath :

AMD admitted the real TDP is 90W; third-party reviews confirmed the R7-1700 is a 90W chip; AMD is bundling the 1700 with a 95W cooler. Links and quotes given in point (6) in the former resume.

adamsleath :

This is an AMD vs AMD issue. Phenom, Bulldozer, Piledriver, ThreadRipper and R5/R3 RyZen satisfy the official TDP. R7 Ryzen violates the offficial TDP. Proofs and explanations in point (7) in the former resume.

P.S.: Correcting mistakes, giving accurate info, and links to reviews is not trolling.

juanrga · Dec 30, 2017

(a)
As explained in the well-known book by Hennessy & Patterson

TDP = sustained power consumption

Second, what is the sustained power consumption? This metric is widely called the thermal design power (TDP), since it determines the cooling requirement.

(b)
TDP is associated to real-life usage of a component. Real-life usage for a CPU implies running real-life workloads.

"Torture test" with Prime95 is not a real-life workload. It is a stress-test, also known as power virus for obvious reasons.

Prime95 uses unending loops of full AVX-code. This doesn't correspond to any real-life workload, where small AVX-vode snippets are inserted within normal x86 code.

The goal of a "torture test" is to stress a CPU beyond the limits. Prime95 is often used for stability studies. If your overclocked CPU is stable during 24H under torture with Prime95, then it is stable with any real-life workload.

(c)
Prime95 doesn't stress RyZen CPUs. One can check this by looking how R7-1800X consumes the same power under ordinary workloads as luxrender than under Prime95. The Stilt also noted this in his review of Zen

Note: Current versions of Prime95 (28.10) do not stress Ryzen CPUs properly. The resulting power consumption is abnormally low, and both Firestarter and Linpack result in significantly higher power consumption.

So one cannot use Prime95 to compare power consumption of Zen and non-Zen CPUs. The result of such comparison is biased, because there are a systematic error.

Using another workload, the Stilt got that the 1800X goes above the 95W. The Silt got similar numbers than the HFR review.

(d)
Many initial reviews of SKL-X used pre-launch BIOS with non-working turbo/parking policies, which increased the power consumption abnormally in some models. As reviews noted "I’m betting that this issue will get worked out with a BIOS update". Indeed, latest BIOS corrected this.

(e)
Anandtech measuring 149W at the package level doesn't imply a violation of the 140W TDP. First, because those 149W were measured during a "torture test", not a real-life workload. Second, because the difference bwtween 149W and 140W is of only 6% which is within the margin of error of the measurement. Anandtech also finds the 16-core i9 to satisfy the 165W TDP. However, the 18-core i9 consumed 190W. Ignoring for a minute this was a torture test, this 15% is too big to be explained by error measurements. So there is something wrong here. It is evident that their BIOS/chip combination is wrong because other reviews didn't find the same than Anandtech found. In fact, several reviews found the 18-core i9 consumes less power than the 16-core i9 thank to using a better binned die.

Anandtech also knows the relation between TDP and power consumption:

A processor with a TDP of 140W should, in general, consume 140W of power (plus some efficiency losses).

A demonstration that all the i9 chips satisfy the official TDPs is found in point (10) in the former resume. The i9-7900X is within the 140W TDP. Whereas the i9-7960X and the i9-7980XE are within the 165W TDP.

goldstone77 · Dec 30, 2017

juanrga :

TDP = sustained power consumption

TDP in the past when concerning CPU's was directly related to power consumption, which is consistent with the first law of thermodynamics. Using a HSF(entropy to equilibrium) is consistent with the second law of thermodynamics. None of which is in question for me. What is in question is specific manufacturers definitions of TDP, which have become marketing terms, and they define as follows:

According to AMD, "TDP is essentially the maximum sustained power a processor can draw with “real world” software while operating under defined temperature and voltage limits."

According to Intel, "TDP represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload."

Pushing all of the Core i9-7900X’s cores with Prime95 or LuxRender propels power consumption to incredible heights. You do get 48 percent more rendering performance in LuxRender, but at the expense of 58 percent-higher power use. This approach has the elegance of a sledgehammer. Then again, if you need speed at any cost, Core i9-7900X is top-notch.

LuxRender, but we can talk about AVX workloads, but that's doesn't count? We just choose to ignore it, right? Because that's not fair to Intel, because it defies their

According to Intel, "TDP represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload."

The Stilt:

The power consumption

All of the power consumption measurements have been made with DCR method. The figures represent the total combined power consumed by the CPU cores (VDDCR_CPU, Plane 1) and the data fabric / the peripherals (VDDCR_SOC, Plane 2). These figures do not include switching or conduction losses.

Peak power (i.e. worst-case) figures were measured during Firestarter FMA/AVX binary execution. On average the resulting power consumption is around 30% higher than the power consumption resulting from any other real world consumer, fully multithreaded workload.

Note: Current versions of Prime95 (28.10) do not stress Ryzen CPUs properly. The resulting power consumption is abnormally low, and both Firestarter and Linpack result in significantly higher power consumption.

Peak Power overclocked to 3.9GHz! Stock had a much lower "Peak Power" of 110.994.

According to AMD, "TDP is essentially the maximum sustained power a processor can draw with “real world” software while operating under defined temperature and voltage limits."

An easter egg

Zeppelin features a highly advanced power management, as stated many times before. Just like Carrizo / Bristol Ridge, which feature a very similar PM, Zeppelin can infact support cTDP as well. cTDP is not officially supported (or available) on any consumer Zeppelin based SKU (AFAIK). The lack of official support is merely a distraction

850 points in Cinebench 15 at 30W is quite telling. Or not telling, but absolutely massive. Zeppelin can reach absolutely monstrous and unseen levels of efficiency, as long as it operates within its ideal frequency range.

Edit: Added quotes.

juanrga · Dec 30, 2017

goldstone77 :

The HFR review uses an AVX-enabled workload to measure TDPs.

goldstone77 :

By "peak power" he doesn't mean peak power in the same sense used in this thread. By peak power he means a worst-case sustained load. His worse-case doesn't have to be the worst-case for everyone else. For him the worst-case was FireStarter, using a custom build because the official version doesn't support RyZen. He got 111W of sustained power, which is the same obtained by HFR with an ordinary x264 workload: "The Silt got similar numbers than the HFR review." So his worst-case is an ordinary case for others.

goldstone77 :

It is worth to mention that CineBench is a special workload that favors RyZen and that he is plotting CB scores vs setting cTPDs, which don't have to agree with real TDPs. Next an efficiency graph with CB

goldstone77 · Dec 30, 2017

juanrga :

By "peak power" he doesn't mean peak power in the same sense used in this thread. By peak power he means a worst-case sustained load. His worse-case doesn't have to be the worst-case for everyone else. For him the worst-case was FireStarter, using a custom build because the official version doesn't support RyZen. He got 111W of sustained power, which is the same obtained by HFR with an ordinary x264 workload: "The Silt got similar numbers than the HFR review." So his worst-case is an ordinary case for others.

In the context of the review as he defines it, is different from what you are trying to use in the quote to support HFR. It doesn't support the ~128W of power consumptions referred to HFR. The graph shows ~111W of "Peak Power" in a stock configuration, which he defines in his test being 30% higher than power consumption resulting from any other real world consumer, fully multithreaded workload. And AVX workload drastically increase Intel power consumption, that's why programs like prime95 test Intel CPU's with it disabled to prevent thermal throttling.

Peak power (i.e. worst-case) figures were measured during Firestarter FMA/AVX binary execution. On average the resulting power consumption is around 30% higher than the power consumption resulting from any other real world consumer, fully multithreaded workload.

The real argument that has spanned miles on this thread comes down to the multiple definitions of TDP. Manufacturers have modified it over the years creating guidelines on how they determine TDP that are vague. There isn't a standard, meaning we don't have a defacto test that is universal approved by these two companies to compare their TDP's with predefined specifications. AMD: “real world” software while operating under defined temperature and voltage limits." Which temperature and voltage? Intel: "average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload." Average watts dissipated on all cores at base frequency that intel-defines as a high-complexity workload. This leaves the power of defining TDP in the hands of the manufacturers. No, amount of arguing over past and present will resolve these definitions as they are, because they can mean almost anything they want it too.

According to AMD, "TDP is essentially the maximum sustained power a processor can draw with “real world” software while operating under defined temperature and voltage limits."

According to Intel, "TDP represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload."

Edit: fixed typos

juanrga · Dec 31, 2017

goldstone77 :

The statistical problem is two-dimensional. There is a temporal dimension and a workload dimension.

In this thread, and in the book by Hennessy & Patterson, peak power is defined in the temporal dimension. I.e., there is some time t_p for which the instantaneous power P measured for some workload w is the maximum recorded in the time series

P(w, t_p) >= P(w, t).

The Stilt is defining peak power in the workload dimension. I.e., there is some workload w_p for which the sustained power <P> = {(1/T) int_0^T P(w,t) dt} is the maximum recorded compared to the rest of workloads

<P(w_p)> >= <P(w)>.

For The Stilt, w_p = "custom build of FireStarter". But w_p can be different for others, because others can use (in fact use) workloads that The Stilt didn't run in his review.

Also if you pay attention to the IPC measurements, you can see that The Stilt got that Kabylake is 30% faster than Zen. No other review got a 30% average gap in the IPC. HFR obtains IPC gap of less than 20% between Zen and Kabylake.

This difference in the measured IPC suggests that The Stilt is using custom workloads that don't stress RyZen enough. This explains why the IPC is lower than in mainstream reviews, and this could explain why The Stilt gets lower power consumption, with his worst case scenario corresponding to an ordinary workload for others, because his custom workloads aren't stressing the Zen CPU.

Also the 128.9W measured by HFR is on the ATX12V rail. This is not the power consumption of the CPU alone. To get the power consumption of the CPU one has to use the 85% factor that accounts for the VRM and other efficiency loses. Using this factor the power consumption of the CPU alone was ~110W, which is close to the 111W measured by The Stilt using FireStarter. From the HFR review (once again):

Indeed with a consumption measured on the ATX12 at 128.9 watts, it is obvious that the consumption of the Ryzen 7 1800X exceeds the 95 watts announced on TDP (Thermal Design Power). Indeed even if we base on a yield of 85% in the power stage of the motherboard, we arrive at almost 110 watts. An estimate confirmed by the internal monitoring of the processor which even indicates 112 watts under x264.

This all has been explained a hundred of times. The 85% factor has been used in multiple posts in this thread to get the sustained power of several CPUs. But now it is all forgotten?

I already answered about AVX workloads. I just mentioned above that HFR is using an AVX workload for all the CPUs Intel and AMD. And I explained in former posts that AVX workloads also drastically increase power consumption in AMD chips. ThreadRipper was used as example. The ThreadRipper 1950X CPU is stressed up to the point it has to underclock cores under the base clock. All this was extensively discussed and proved, with links and graph given

goldstone77 :

As mentioned a hundred of times, this is not AMD vs Intel. This is AMD vs AMD.

As demonstrated before, Phenom, Bulldozer, Piledriver, ThreadRipper, and R3/R5 RyZen models work within the official TDP. Measurements and graphs given before. R7 RyZen models don't work withing the official TDP. Measurements and graphs given before. So it is not about AMD vs Intel.

I am discussing AMD in a thread about AMD products and some of you insist on mentioning Intel again and again.

We could for a minute believe that AMD changed the definition of TDP for Zen and that it doesn't matter that Phenom, Bulldozer, and Piledriver chips run within the official TDP. But we would be fooling ourselves because some Zen products violate the TDP, whereas other TDP products don't.

The R7 RyZen models violate the official TDP. The R3/R5 RyXZen models and the ThreadRipper models don't. So it is not about the definition used. AMD has not changed the definition from the R7 to the R5, neither from the R7 to ThreadRipper. I also explained why the R7 models violate the TDP and why the R5 or ThreadRippers don't.

The first law of thermo applies to every CPU. If the R7-1800X is consuming a sustained power of 112W then it is not a 95W chip. If the R7-1700 is consuming a sustained power of 90W, then it is not a 65W chip. Period.

I think there is no more to say here. I am refuting the same arguments again and again. The discussion goes in circles.

Gon Freecss · Jan 1, 2018

Yeah, you proved your point. They're going to side step the argument and go on about how TDP is a marketing scheme. AMD just wanted to maximize performance when Ryzen 7 launched.

adamsleath · Jan 1, 2018

juanrga :

you arent correcting mistakes. your information is inaccurate inasmuch as they are inconsistent with all data sets.

Intel 'violates' their nominal TDP's ..that whole issue is just a bore right now.

just a mass of intel spam in an AMD thread.
------
"As mentioned a hundred of times, this is not AMD vs Intel. This is AMD vs AMD." a statement as meaningless as it is annoying. just more intel spam.

"I think there is no more to say here. I am refuting the same arguments again and again. The discussion goes in circles."
because it is trivial and boring? and the points were already made many posts previously?

you are still deluding yourself and ignoring tests that show intel violating their tdp limits - using intel's own definition.
that is the interminably annoying slant to your posts. no wonder youve been banned from other forums.

youre also blinded by your own apparent need to be correct, even when youve been proven to be incorrect; this is a weakness that is easily addressed, but probably wont be due to bias and denial.

AMD's Future Chips & SoC's: News, Info & Rumours.

Distinguished

Distinguished

Distinguished

Distinguished

Honorable

Splendid

Distinguished

Distinguished

Honorable

Distinguished

Reputable

Glorious

Distinguished

Distinguished

Reputable

Distinguished

Honorable

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Reputable

Honorable

Share this page