Podcast #478 - Windows on ARM, Intel 10nm rumors, and more!

Subject: General Tech | December 7, 2017 - 01:45 PM |
Tagged: xfx, Vega, Raspberry Pi, radeon, qualcomm, nicehash, Intel, IME, GTX 1070Ti, gddr6, evga, Elgato, dell, coolermaster, cluster, asus, arm, amd, AM4, Adrenalin Edition, 4k60, 10nm

PC Perspective Podcast #478 - 12/07/17

Join us for discussion on Windows on ARM, Intel 10nm rumors, and more!

You can subscribe to us through iTunes and you can still access it directly through the RSS page HERE.

The URL for the podcast is: http://pcper.com/podcast - Share with your friends!

Hosts: Ryan Shrout, Josh Walrath, Jeremy Hellstrom, Allyn Malventano, Jim Tanous

Peanut Gallery: Alex Lustenberg

Program length: 1:39:42

Podcast topics of discussion:
  1. Week in Review:
  2. News items of interest:
  3. Closing/outro

Source:

December 7, 2017 | 03:11 PM - Posted by Power (not verified)

"Whose is longer, AMD's Ryzen Mobile or Intel's Kaby Lake-R?"

Seems you need to read TR some more:
https://techreport.com/news/32920/report-hp-envy-x360-battery-life-drags...

December 7, 2017 | 06:07 PM - Posted by Jeremy Hellstrom

You mean Laptop Mag, seeing as how that is a news post about another publications testing that The Tech Report linked to after publishing the story I found interesting?

December 7, 2017 | 06:53 PM - Posted by rl (not verified)

Question is, why didn't you mention it?
Weren't you aware or did you choose to focus on that one test?

Because that first test makes it look like Ryzen Mobile sucks... a lot of battery while Kaby Lake sips it. And the other test seems to show that it's HP that really sucks at making decent laptops with good battery life in general. *shrug

December 7, 2017 | 08:21 PM - Posted by OnlyFoolsMakeRushJudgementsThisEarlyInTheGame (not verified)

It's still to early to tell absent any other Ryzen/Raven Ridge SKUs to test in other OEM Laptop SKUs. So AMD's Raven Ridge APUs are going to need longer before any informed results can be obtained. And any final differences between AMD and Intel APU/SOC SKUs with Integrated graphics could be down to how much power the laptop motherboards use and the respective MB's related chipsets and other factors such as Thunderbolt controllers and such, and AMD's use of more powerful graphics in its APUs will use more power.

The thing I'd expect maybe is that AMD's graphics being a little more performant than Intel's integrated graphics in that AMD's and Nvidia's grahics tend to have higher shader counts than Intel's graphics. So AMD's integrated Vega 8/nCU graphics is already ahead Intel's graphics and APUs using Vega integrated 10/nCU graphics is going to take more power but really add to AMD's higher lead in integrated graphics performance.

Most people looking at AMD's APUs are looking at that Vega graphics more than they are looking at battery life and I'm more interested in what the ASUS/other laptops using the Ryzen 7 2700U and desktop Raven Ridge APUs will do. And yes I really want a laptop SKU that makes use of a Desktop Raven/Ridge APU and I'll use the laptop plugged in as a desktop replacement laptop and battery life in not a concern.

What Tech sites need to be foucsing on more is asking ASUS, who already sell a laptop with a desktop Ryzen 7 1700 inside, is will ASUS be doing the same for some desktop Raven Ridge APU SKUs with Vega integrated 11/nCU based graphics at say 45-65 watts. Because that opens up the market for some affordable laptop SKUs that really can perform without the need for any extra discrete mobile GPUs and that is somthing that would be interested in purchasing.

All this laptop battery life testing has too many unknowns and laptop power uasge is more affected by what the laptop OEM includes on the laptop and that includes battery capacity and other motherboatd connectivity options that play a bigger role in battery life in addition to the CPU/APU makers' respective MB chipsets.

If you look at Intel's integrated graphics performance without the help of Nvidia or AMD it's really Intel's integrated graphics that sucks and folks looking to AMD's Raven Ridge care about the APU's Graphics performance more than the APU's battery life. AMD's Raven Ridge APU true value is in the graphics performance at a lower cost than that poor graphics that Intel trys and foist on to the market.

If you really want low power and do not care about anything else that maybe look at the ARM based SKUs that can now run some windows/win32 applications and forget about x86.

I still hope that AMD is not foolish enough to throw oth that custom ARM/K12 projsct IP and tru and become to much like Intel and be too dependent on one ISA like x86. AMD has a big chance with that K12 custom ARM design to throw off that x86 CISC ISA power usage intrensic disatvantage and get to market some lower power using K12 ARM/Vega APUs that will compete better in the new windows on ARM market for those that want even longer battery life. AMD's biggest mistake will be not to continue to move away from any dependence on only the x86 ISA. AMD needs to go full on with some custom ARM/K12 based APUs for tablet/hybrid designs where battery life is the most concern.

Only Fools Make Rush Judgments this early in the game, and any websites making judgments this early are most likely not Fools! As it's more likely that those websites are being incentivized by someone to spin things the way they are counter to logic that says it's too soon to pass judgment for lack available testing samples or even some unknown variables. Show me a website that comes to a rash judgment about a new product first time to market that later has that website having to retract that judgement at a later time when some peers at other websites get different results, or reveal some new unknowns! And that can only cause suspicion of that website's rash judgment being incentivized in some manner.

December 7, 2017 | 08:32 PM - Posted by Anony mouse (not verified)

They might not been aware of it but they should correct it for their audience written and podcast.

If PCPerspective is relying on other publications for information on their test they should also report on follow ups those publications do.

December 8, 2017 | 06:31 PM - Posted by Jeremy Hellstrom

Ah yes, the old argument that one must be aware of 100% of everything on the internet to be allowed to say anything. 

Missing one single article or follow up invalidates everything we have ever or will ever say?

December 8, 2017 | 08:40 PM - Posted by Way neJetSki (not verified)

hahaha Yes exactly

December 8, 2017 | 12:03 AM - Posted by NoROPsNoFPS (not verified)

Why would Nvidia announce their Titan V and report only the Texture unit count and texture fill rate and not include the ROP counts and Pixel Fill rate.

320 TMUs on Titan V but what about ROPs and the Pixel fill rate? And can the tensor cores be used for any graphics processing tasks. Hardocp lists the taxture fill rate at 384 Gtextels/s. But where are the ROP counts and pixel fill rates?

December 8, 2017 | 12:07 AM - Posted by NoROPsNoFPS (not verified)

P.S. the Vega 64 has a texture fill rate of 395.8 GTexel/s, so what is up with that Titan V texture rate.

December 8, 2017 | 01:17 PM - Posted by partizan (not verified)

Yeah, it is a bad test - they are using 2 totaly different HW laptops - the swift has DDR3LP on board while the HP has 2 memory sticks, hp has a digitizer which doesnt get dissabled by disabling the display, hp also has additional HW like SD card reader which the acer doesnt have. By the way they updated their conclusion on tech report. So please stop spreading half truths w/o researching stuff and sh*ting on AMD based on a 7 days old news, which are already corrected for 4 days - https://techreport.com/news/32920/report-hp-envy-x360-battery-life-drags...

December 11, 2017 | 05:03 PM - Posted by BillR (not verified)

I'd bet that largest factor in getting fixes out for the Intel management engine issue is mostly on the vendor validation front. Each vendor has to take the build from Intel then build and validate the BIOS for each and every system they made. For older computers add the problem of finding years old platforms, getting what are probably early engineering builds running reliably (if they still have them), recreating old build environments accurately, then building/testing the BIOS. Might be a problem for any PC over a couple of years old.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.