PC Perspective Podcast #400 – 05/19/2016
Join us this week as we discuss the GTX 1080 performance and features, official specifications of the GTX 1070, new Polaris specification rumors, ARM's 10nm chip test 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: https://pcper.com/podcast – Share with your friends!
- iTunes – Subscribe to the podcast directly through the iTunes Store (audio only)
- Google Play – Subscribe to our audio podcast directly through Google Play!
- RSS – Subscribe through your regular RSS reader (audio only)
- MP3 – Direct download link to the MP3 file
Hosts: Ryan Shrout, Jeremy Hellstrom, Josh Walrath, and Allyn Malventano
Program length: 1:03:21
-
Week in Review:
-
-
News items of interest:
-
1:04:30 Hardware/Software Picks of the Week
-
Ryan: DOOM
-
Jeremy: Triple your phase change memory
-
Allyn: Sony RX10M3
-
-
Closing/outro
Subscribe to the PC Perspective YouTube Channel for more videos, reviews and podcasts!!
“Trying Out The Experimental
“Trying Out The Experimental AMDGPU OverDrive Overclocking With The R9 Fury”
TuxBoost!
https://www.phoronix.com/scan.php?page=news_item&px=AMDGPU-OverDrive-R9-Fury
Neat!
Neat!
As for Async Compute. You
As for Async Compute. You forgetting the biggest market the PC Gamers don’t want to ever acknowledge.
The consoles are putting it into a wider use both XB1 & PS4. Tomb Raider is a good example where the console version had more Async and during the port GameWorks replaced certain things done in Async Compute to Nvidia GameWork middle-ware features.
Also, saying that it is a
Also, saying that it is a small performance difference is missing the point. It seems like a very useful feature in many instances. The Ashes of the Singularity developers seem to have good reasons to use it. With Nvidia essentially not implementing it (yet), developers now have to do two different versions, one for consoles and AMD cards and one for Nvidia cards. While Nvidia has a large share of the PC market, AMD has both major consoles.
Why didn’t Nvidia implement it sooner? I would guess that it is a similar situation to when Intel refused to support DDR with the P4 for a while. They didn’t want to support it because it would give AMD and advantage. The AMD K7 used DDR and Intel was using Rambus. One of my friends bought a P4 in the time Intel was not supporting DDR, and because of the expense of Rambus, he ended up with a P4 with SDR. It was probably a performance downgrade from his old machine; the P4 performed terrible with SDR ram. Such practices hurt consumers.
Nvidia was fine with DX11, since they put so much software development into it. AMD put resources into developing Mantle instead, which I consider a good thing. AMD did a huge amount to push the industry forward by doing so. Intel wouldn’t have wanted DX12 either, since removing the reliance on single thread performance would help AMD processors of the time. With Zen, they shouldn’t need the help, but we were obviously long overdue for an properly multi-threaded graphics API anyway. Microsoft pretty much had to push DX12 for the Xbox1 with the 8 low power core configuration.
It am not a fanboy for either company; I have had machines with both AMD and Nvidia cards. I even had a Matrox G200 back in the 90s of so. I tried to boot the old k6-300/Matrox G200 machine up a while back, and I didn’t get any video, so perhaps my Matrox card has failed. I find it hard to support Nvidia at the moment, even though my main machine is currently Intel and Nvidia based. I see way to many things from Nvidia that are monopolistic and/or anti-competitive practices. Any company in a dominant position will generally pull such tactics, but I will not support such things. I also tend to keep my hardware a long time. I suspect an AMD 390x will still be a capable card for some future DX12 games. I suspect current Nvidia cards may not do so well. Nvidia card owners can count on Nvidia’s current market share to get them custom optimization for quite a while yet though.
I’m a big fan of getting more
I’m a big fan of getting more of both the gaming graphics/gaming compute and the graphics rendering compute(Namely Ray Tracing) done as much as possible on the GPU’s hardware. So the more in hardware CPU like async-compute features the GPUs from both AMD and Nvidia/others get in their respective GPUs’ hardware the better it will be for games and non gaming graphics done on the GPU!
Imagination Technologies is getting options for dedicated Ray Tracing rendering units on its Mobile PowerVR variants and some Graphics tablets that have that PowerVR Ray Tracing units on their SOCs’ Integrated GPUs will make for some great graphics oriented tablets for graphics workloads, as well as more realistic graphics/gaming lighting/shadow/AO effects without using the CPU at all.
There is even OpenCL/Cuda based ray tracing middleware(from both AMD/Nvidia) for accelerating Ray Tracing calculations on the current and older GPU generations’ graphics that can speed up more ray tracing done on the GPU’s shader/other units to assist in accelerating ray tracing workloads more on the GPU’s core assets. But I’m hoping that all GPUs will start getting specialized Ray Tracing units built into the GPUs hardware because that is the most efficient way to do Ray Tracing on the GPU.
The CPU, even Intel’s Xeon SKUs, takes forever doing the ray tracing calculations that could be much better done on a GPUs specialized units that outnumber the CPUs SIMD/other units counts by many orders of magnitude. And the more Gaming graphics/Gaming compute that can run on the GPU without any CPU help the less there will be any of the CPU to GPU over PCIe latency issues that can really mess up the VR gaming experience. So the more CPU like functionality that AMD engineers into its GCN GPU hardware ACE units the less dependency there will be on the CPU for most all gaming/graphics workloads, ditto for Nvidia and its future GPU variants.
The Mobile market is sure going to be using the Vulkan API to its fullest extent, and most of the the mobile markets CPU/APU/GPU HSA foundation members will be trying to get as much HSA 1.0(and newer) compliance to do as much work on the GPU as possible, and across as many CPU cores also. The GPU is gradually gaining more and more of the functionality that was traditionally in the domain of the CPU, so keep your eyes on AMD’s RTG, as they will be adding more of that CPU like functionality into the ACE units, for both the VR gaming market and the HPC/Server/Workstation markets and the exascale market that is currently receiving plenty of government funding from the government exascale computing initiative!
The latest Bristol Ridge APUs
The latest Bristol Ridge APUs based off of a tweaked Excavator architecture are Set to launch during Computex, so in addition to the Polaris GPU launch the Bristol Ridge APUs/AM4 motherboards/others there will probably be new AMD Laptop/PC design win news. So what about the previous Carrizo SKUs being thermally locked down(15 watts and very little 35 watt options for parts with 35 watt capabilities) by the laptop OEMs, that single memory channel gimping, and will there be more laptops with DDR4 options, now that the Bristol Ridge APU SKUs will support DDR4 memory options.
I hope that AMD can use any of their future Zen’s platform’s strengths/options and future laptop/desktop market options with the OEMs as some form of AMD clout to convince mostly the laptop OEMs to not gimp down the AMD Bristol Ridge laptop SKUs so much. Because I’m still looking for a new laptop that uses the Bristol Ridge Excavator FX[highest end/wherever #] laptop SKU with the best graphics that is not gimped down by the laptop OEMs. AMD needs to be asked the hard questions at Computex and any interviewers should not let AMD off the hook so easily if AMD tries to avoid the hard questions about all the OEM laptop SKU gimping. Someone should also ask AMD for a time-line on when the first Zen/Polaris laptop SKUs will be expected, if AMD decides to talk a little bit more about their Zen based systems.
AMD also needs to be asked if they plan on getting, or are getting, any Linux OS based laptop OEM design wins because there will be many laptop users interested in that the closer to 2020 it gets. Hopefully the Zen/Polaris laptop SKUs will prove to be too attractive for the Linux OS based OEMs to pass up when they arrive and there will be many Zen/Polaris APU designs wins among the Linux OS based laptop OEMs, and that includes some Polaris discrete GPU options to boost the Zen/Polaris APU integrated graphics ACE unit total counts on Laptop SKUs, both Linux/other OS based laptop SKUs. I’d love to see Apple commission some custom Zen/Polaris APU on an interposer designs from AMD and get that and HBM2 into Apple’s Macbook Pro lines of laptops, because most Apple laptop hardware will run a Linux OS build also.
another overpriced product.
another overpriced product. In india it is priced at 1000 dollars thats two much. i ask both amd and nvidia to stop overpricing their product in india.
Nobody is forcing you to buy
Nobody is forcing you to buy overpriced products… it is the bid and ask market!
i think your in the wrong
i think your in the wrong place to complain about pricing of amd & nvidia products 🙂
Thanks for explaining (a lot)
Thanks for explaining (a lot) ASync compute, I’m waiting on those games that will be able to take advantage of this technology (Nvidia or AMD), at the end what matters is the result on both sides.
The problem might be with developers, I don’t see them implementing both technology (like always games and partners…), hoping it won’t break games for the other. Because I think is sad to see such a drop in performances between games if you use AMD or Nvidia.
On the gpu side I’m really waiting on the gtx 1070 and the polaris 10 depending on performance and price but I’m guessing the 1070 will be a better deal.
Keep up the good work guys!
“New AMD Summit Ridge CPU
“New AMD Summit Ridge CPU Dieshot
Exclusive: Calming … zen gardens of silicon.”
http://semiaccurate.com/2016/05/22/38688/