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

Page 71 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.


It is not a single test. Ian claims he got 22% IPC improvements in other benches, which is NONSENSE. And their gaming numbers are NONSENSE, and their claims about the microarchitecture are NONSENSE. In the cache section Ian claims L1 latency is reduced by one cycle. This is false. 1800X vs 2700X both at same clocks at same memory show the same L1 latency in nanosecond, because the cycles are the same; the cache is unchanged.
 
Ok, so I thought of a simple way to stop the "TDP is Power Consumption" misconception once and for all with a simple paragraph:

When the CPU consumes power from circuitry, it needs to convert the voltages and amperage into what it can handle, you there's an implicit increase in "waste" (heat) when having to transform that in the CPU itself (Intel and AMD have the regulators inside the SoC; NDA documents, unfortunately). After that is done and the energy is going into the CPU, then you start actually working: the transistors generate heat due to resistance and generates thermal transfer to the closest suitable transfer point. What this point is, could might well be the bottom of the MoBo or the HS plastered to the CPU. Then you have the efficiency of the transfer itself. We all know that anything that has a resistance gets affected by heat; more heat, usually generates more resistance and in turn you need more power to actually operate. Interestingly enough, this is not related to temperature. You can operate at a nominal 60°C and have a completely different power consumption with the same CPU family, but still work within "normal" limits of the materials that compose the CPU/transistors and the paste/solder used to help the heat be transferred out.

I think I should make a drawing out of this and post it...

I really really hope this helps in understanding that TDP != Power Consumption. Yes, they are related, but by no means the same thing.

Cheers!

EDIT:


This is what Ian wrote about the latency:

It is interesting that in the official slide deck AMD quotes latency measured as time, although in private conversations in our briefing it was discussed in terms of clock cycles. Ultimately latency measured as time can take advantage of other internal enhancements; however a pure engineer prefers to discuss clock cycles.
Click to expand...

You're being really unfair, I insist.

I'll stop here though; I don't want another ban.
 
MD R7 2700 & 2700X Review: Streaming Benchmarks, Memory Scaling, & Volt-Frequency Performance
By Steve Burke Published April 19, 2018 at 9:00 am

Power Testing – AMD R7 2700(X) Power Draw
This is where it gets a bit interesting for AMD. AMD’s version of “TDP” isn’t comparable to Intel’s. You can’t point to a 95W TDP on the 8700K and compare it to a 95W TDP on the 1700X. The numbers are reached in different ways.

For AMD, TDP is calculated by subtracting 42 from 61.8 and dividing by 0.189. Those numbers are derived from the following: AMD claims that the “optimal” tCase temperature is 61.8 degrees, hence 61.8. AMD also says the optimal ambient temperature at entrance to the heatsink fan is 42 degrees at the inlet. They also say that the minimum degrees per Watt rating of a heatsink to achieve rated performance should be 0.189 thermal resistance.
Click to expand...

http://www.tomshardware.com/forum/id-3341285/amd-naples-server-cpu-info-rumours/page-34.html#20900161
 
DwPWcLa.png

https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/page-72#post-39391302
Personally, I think that AMD should have rated these CPUs for 140W TDP instead of the 105W rating they ended up with. The R7 2700X is the first AMD CPU I’ve ever witnessed to consume more power than its advertised power rating. And honestly, I don’t like the fact one bit. Similar practices are being exercised on Ryzen Mobile line-up, however with one major difference: The higher than advertised (e.g. 25W boost on 15W rated SKUs) power limit is not sustainable, but instead a short-term limit like on Intel CPUs. The way I see it, either these CPUs should have been rated for 140W from the get-go, or alternatively the 141.75W power limit should have been a short-term one and the advertised 105W figure a sustained one.

According to AMD, the TDP is determined as follows: tCaseMax – tAmbientMax / ϴca (the minimum thermal resistance of the cooling element, °C/W). For the 105W rated SKUs these limits are following: tCaseMax = 61.8°C, tAmbient = 42.0°C and ϴca 0.189°C/W. Regardless how the advertised power rating has been established, it doesn’t change the fact that the actual power consumption on these parts is higher than advertised and more importantly how the consumers generally perceive and compare the advertised power ratings of the different CPUs.

It is not uncommon for a modern CPU to temporarily exceed its rated TDP, as most of the infrastructure definitions from both AMD and Intel include such functionality by the design. For example, recent Intel CPUs, such as Skylake and newer have the boost (PL2) power limit set 25% higher than their rated TDP (PL1). However, the raised boost limit is only available for a thermally insignificant amount of time (1 second on Intel by the specification).

Since 105W TDP rated Pinnacle Ridge CPUs are allowed to sustain >= 141.75W of power draw, and more importantly because at stock they do consume significantly more than the rated 105W even in real world multithreaded workloads, their advertised power rating in my opinion is not entirely fair and might end up misleading the consumers. The measured sustained power consumption for a stock 2700X was 127.64W (132W peak) during X264 encoding and 142.52W (146.5W peak) during Prime95 28.10. In comparison, a stock i9-7960X CPU with its power limit reduced from the default 165 / 206W to 140 / 175W (PL1, PL2) sustained 139.82W power draw and had a peak draw of 168W in the very same X264 workload. All of the stated power figures are based on DCR (current over inductor) measurements and therefore external conversion (VRM, PSU) losses are not included in them.

Despite the rant about the power consumption, at least to me the issue is more of an ideological one than a practical one. The 2700X SKU ships with the biggest factory heatsink (with 0.170 ϴca °C/W according to AMD) solution the industry has ever seen, so cooling wise it is pretty much irrelevant if the power dissipation of the CPU has been slightly understated or not. Regardless, in my opinion both AMD and Intel should clearly state both the sustainable and the boost / peak power figures for their products, preferably right at the product page at their website. Currently the information is either completely unavailable to the consumers (i.e. NDA required) or alternatively rather hard to come by.
Click to expand...
Skylake and newer have the boost (PL2) power limit set 25% higher than their rated TDP (PL1). However, the raised boost limit is only available for a thermally insignificant amount of time (1 second on Intel by the specification).
Click to expand...

Can we say that better CPU coolers lead to higher sustained boost for Intel processors? Which would in turn lead to greater power consumption!
 
GamerNexus' video for the Streaming comparison of the 2600X and the i5 8600K is quite telling:
[video="https://www.youtube.com/watch?v=GDggr3kt96Q"]https://www.youtube.com/watch?v=GDggr3kt96Q[/video]

I don't understand how it can be so [strike]un-playable[/strike] un-watchable when I can stream with little effort using my i7 2700K. Given generational and architectural changes, they should be equal or the i5 8600K edge out my CPU. I can't at 1080p without frame drop, but it's never *that* bad either.

Cheers!

EDIT: Typo.
 

This seams to be an Intel CPU issue overall. I think it's important to note the settings used(quality) is what is causing the Intel CPUs to completely fail, but then again that is part of the test. Linus exposed some of these quality issues in his video last year!

Ryzen is THE BEST CPU for Game Streaming? - $h!t Manufacturers Say Ep. 2
Linus Tech Tips
Published on Apr 6, 2017
Is Ryzen REALLY the best consumer CPU option for video encoding and game streaming? Let's find out!
[video="https://www.youtube.com/watch?v=jludqTnPpnU"][/video]
 
I did not see that video from LTT, thanks.

Toms did an excellent article comparison as well, but they didn't display streamed video (or how it would look). It is quite different to see the data and actually seeing how it's in the stream itself. I guess I'm more shocked of my own lack of imagination/projection on the data, haha.

I do make tests when I stream to make sure it's fluid (that's how I've nailed profiles for each game I like to Stream), but like I said, it's never been that bad. I wonder if the patches have anything to do with it as well?

Cheers!
 
My friends and I stream on steam all the time when one of us gets a new game or just to BS(using 2500k, 3770k, 7700K, 8700K, and 9590 processors). It's never that bad, but it's really not that great either to be honest. Also a good comparison is Twitch live streams. Most tell you what hardware they are using, and some use a separate machine to handle the streaming. Those are the best quality streams.
 


Best my 1700 can do is 1.15V at 3.6Ghz and it craps out at 3.85Ghz at 1.375V. Newer chips are way better but the 1000 series did get better with the later production. With a lot of people getting 3.8Ghz at only 1.25V or so
 


Now lets hope they don't troll us with 4C/4T Ryzen 3 2000 series chips(without a igpu)

One thing i like is Amd is packaging heatsinks with the X models this time around was so dumb that they didn't do that its another 20$ advantage for a lot of buyers compared to Intel.
 

Just going to drop this here.

TDP != Power Consumption. I have said that probably close to one thousand times on this forum. Stop trying to mistake the two for being the same.
 


It's a bit more, actually. The Wraith actually is just a tad worse than the 212 EVO (the 95W rated version), so it's really good for a boxed cooling solution: https://www.anandtech.com/show/10500/stock-cooler-roundup-intel-amd-vs-evo-212

Here's Toms: https://www.tomshardware.com/reviews/amd-wraith-cpu-cooler,4450.html

There's one caveat though (thanks to randomizer): it seems the new ones have worse acoustics according to the latest review using Prism here at Toms.

But yes, I agree. Bundling the Wraith with the X models was indeed a clever move. The value proposition increases greatly for non-OC'ers or people that can't or won't buy a cooling solution right away (I've heard and seen it being a common thing, weirdly enough).

Cheers!
 


I know i posted links as well TDP does not equal power consumption linustechtips, wendell from levelone techs, Amd and Intel all state this.

BUT i do think they need to somehow state the power consumption on CPU's but whatever i guess
 
Monday April 23, 2018
MSI X370 Motherboard UEFI Support for Precision Boost 2
I reached out to MSI last week to see when and if it would be updating its X370 chipset motherboards to support Precision Boost 2, that we found to work so good on our water cooled Ryzen 2700X, and it tells us that the support is already in all its X370 EFI. That seems like it happened fairly quickly, but I guess that is a good thing.
Click to expand...
Posted by Kyle 8:08 AM (CDT)
https://www.hardocp.com/news/2018/04/23/msi_x370_motherboard_uefi_support_for_precision_boost_2

Interesting
 
Audacity MP3 Encode
audacity.png

MP3 Encoding with Audacity using the LAME plugin is a media creation workload that depends on single-thread performance. We see a stark performance difference here between Zen+ and Coffee Lake. The Ryzen 2000-series parts are around 30% slower than their Intel counterparts in this test.

The new Zen+ parts provide an 8% improvement in encoding time over the previous Zen CPUs.

Editor's Note: Shortly before the publication of this review, AMD contacted us about improvements to the LAME plugin for Audacity. We tested with the LAME 3.99.3 plugin, which has been the widely available plugin for Audacity for many years and is based on a LAME release from 2011.

AMD has worked with the developer of this plugin, and as a result, there is now a newer version (3.100). This plugin is based on a more modern release of LAME from 2017 and is compiled with a modern toolchain. Improvements include a move to native x64 binaries and SSE2 support. This new plugin has been officially released by the developer on their website. Here's a quick preview of testing on this new plugin:
Click to expand...
audacity-new.png

As you can see, while AMD still loses this performance comparison, the gap is now around 3% instead of 30%. Additionally, the performance of both the AMD and Intel processors improved. This is a more accurate reflection of relative performance, and we will be using the LAME 3.100 plugin in all CPU testing going forward.
Click to expand...
It's amazing the improvement you get when programs are adapted for Ryzen! Now just 3% instead of 30%!

https://www.pcper.com/reviews/Processors/Ryzen-7-2700X-and-Ryzen-5-2600X-Review-Zen-Matures/Media-Encoding-and-Rendering
 


Should probably emphasize this part of the article:

This plugin is based on a more modern release of LAME from 2017 and is compiled with a modern toolchain. Improvements include a move to native x64 binaries and SSE2 support.
Click to expand...
 


No, TDP = Thermal Design Power

In other words, a transistor has an efficiency rating. What that means is that every volt that crosses that transistor loses some energy in the form of heat. This heat loss is rated in watts, and that heat loss is what TDP equates to being.

Essentially, TDP tells you how effective a cooling solution you need based on stock settings and performance across a broad range of varied tasks.

You will note, power consumption is measured in watts, but Watts are a measurement of power transfer. A CPU does not actually use Watts. A CPU uses a connection based on a specific voltage, which you can adjust in your BIOS, and a volt is measured based on resistance in circuit. Additionally, the power cable to your motherboard uses a set voltage to apply power to your MB and the CPU socket, the CPU itself has a dedicated power connector with a more tightly controlled voltage.

Watts are not expressed anywhere in a CPU diagram, nor are they expressed anywhere in your manuals or literature that came with your PC hardware. Why? Because Watts measure transfer of energy, and in this case, that energy is transferred to the heatsink in the form of heat, because it is energy lost over a transistor.

One day, we may not define CPUs by TDP, but by power consumption; however, that will be the day that super conductors at room temperature become a reality. Why? Because super conductors can transfer energy near enough to 100% efficiency that there will be no heat loss.

In the mean time, please, for the love of whatever deity (or lack there of) you believe in, stop confounding Watts and Volts. It makes this engineer ready to rip the grey hair I have left out of my head.
 

Nope TDP does not relate to power consumption it has to deal with wasted energy if anything i mean think about man

https://www.intel.com/content/dam/doc/white-paper/resources-xeon-measuring-processor-power-paper.pdf

Lets say a CPU uses 100 watts at the wall but do to thermal conductivity the TDP is only 65 watts that means the TDP is 65 watts.

Again TDP does not equal power consumption i say this as a non fanboy everyone at this site should know i'm no fanboy of anything and really man i know you aren't either!



If you want i can create products that use 100 watts of energy but have a TDP of 80 watts or lower.

https://www.youtube.com/watch?v=yDWO177BjZY
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom