What kind of CPU are do you think we are getting for next gen? poll

We do are think we are getting are are.

Doubt they’ll use desktop parts unless they have some serious cooling solution. My PS4 already sounds like an airport runway when I play games, not too mention consoles rely more on gpu than cpu so I expect Sony and Ms to continue this trend. I think the cpu bottleneck on consoles won’t be as bad as before considering the performance jump from Jaguar to Zen.

:P

Doesn't matter. AMD's CPU and GPU's are shit on PC that cause more troubles than fun, will perform even worse in consoles - underclocked because of cooling and power consumption...

16 core zen3 at 6ghz

lolwut, many PC gamers are rocking a Zen CPU these days.

Intel's next CPUs are aimed at laptops.

Low per-core-performance is a big issue in PC games. Just because you have good performance in a game that can benefit from more threads, doesn't mean you have shit performance that's utilizing just 2 or 4.

@WESTBLADE: Zen CPUs aren’t that bad, come on. At 144hz the issues show, but if you’re playing 4K/60 or 1440p/60 Ryzen is perfectly fine.

Consoles aren’t going past 60fps, so any respectable Zen 2 equivalent will do.

Yeah but 15% IPC increase bro.

Whatever AMD will offer will be just as good for PC gamers and at half the price than what Intel is offering.

Intel are the ones up shit creek until 2021, but the Intel fanboys are always in denial.

TBH, the newly announced Ryzen 9 3900X looks promising (might be my next build), but i personaly have such a bad history with AMD's CPU's and GPU's that i'm still bitter and hold the grudge til this day + the fact i was making a ''casual PC'' build for my brother (R5 2600X + GTX 1660 + 16GB RAM) and got a fried mobo and i have to travel tomorrow across my whole city and deal with incredibly arrogant staff to get a replacement...

Basically, first AMD build in a very long time and already bad experience... So i'm sorry for my bitterness.

@WESTBLADE: That sounds like it was the board manufacturer's bad, not AMD.

To be fair, I'll probably just stick it out until Intel get their shit together. I probably do not need to change a mobo, cpu and ram, the 8400 should hold out, I'm still contemplating whether to shove a 9700K or even 9900K into it and call it a day. I do like new things though, depends on price to be honest.

Probably should get rid of that wired article, they shot their credibility a while ago. This one explains fairly well.

https://www.extremetech.com/gaming/290209-no-the-playstation-5-doesnt-use-amds-ryzen-3600g

the new AMD CPUs sound really promising.

not sure i'll be upgrading from my i7-6700k for a while though.

I know the mobo is to blame (hope it didn't do something to the CPU though, because i can't test it out due to the lack of any other AM4 socket mobo), of course it could happen with an Intel socket mobo too, but the irony of making a rig with AMD CPU after like 5 years (meanwhile made approx. 10 Intel builds for family members) and instantly running into an issue... XD
I hope you understood what i meant now - had a bad history with AMD CPU's and GPU's (including ATi before the AMD buyout) and relatively zero to none with Intel + NVIDIA... I have a shit luck, but man...

We will see, I think an 8 core 16 thread Zen 2 based cpu with a boost clock of 3.2ghz, that should keep the TDP below 50w.

@WESTBLADE: you must not have been keeping up with what’s going on in the PC world. Zen is the real deal and more and more people are switching to AMD cpu’s, Nvidia gpus still have an edge for now but once Navi is released we’ll see if they can maintain that. AMD has taken “bang for your buck” to a whole new level.

we been at this for how many years now? and it's supposedly about equal to rtx2070 while coming a year later?

Yeah, pretty much on the GPU front. How magically would the new AMD GPU fix the older and newer even recent games that are downright broken because devs don't give a shit about AMD GPU's? AMD as of now has a ~15% marketshare and NVIDIA ~75%. I'm not gonna sacrifice many of my favorite games or deal with major issues, just to save a few bucks.

I be think get we are 10ghz Intel Celeron one core.

I bet the Navi GPUs suck. They are kind of fooling you with this new architecture bull, it is literally a revision, like Maxwell to Pascal (Or was it the other way round).

PS5 is getting the best hardware.

Xbox Whatever will be botched beyond belief. Probably a suped up Xbox One XL

@ajstyles: The best mid range hardware you can get.

@SecretPolice said:

@phbz said:

We do are think we are getting are are.

:P

My bad but yeah, perfect gif.

@ajstyles:

xbox one is the only time that microsoft an inferior console in this generation while the xbox 360 and the first one wasn't .

to be honest with what Microsoft have done with the ONE X , I Believe they will be making the most powerful console in the next gen war .

again i am talking in terms of hardware not software .

I hope MS does the strategy of having the cheapest SKU with reasonable next gen h/w and targets and the most powerful and expensive one.

Masses will purchases the cheaper unit but they will win the internet proxy war through comparisons and weaker unit xbox owners will feel a tribal win.

whatever it is i'm sure it will look so cute! in your little iddy biddy console. im sure it might perform pretty decent too. maybe full potato instead of half.

I may switch to Intel Ice Lake since mobile Ryzen's RX Vega IGP drivers are recent two drivers caused black screen death.

For my Ryzen's mobile RX Vega 8

All 2017 drivers are broken e.g blue screen of death when playing youtube video. Can't deploy Ryzen 2-in1 laptop for mass fleet usage.

All 2018 drivers but two drivers are semi-broken e.g. stuck to a single resolution.

All May 2019 drivers are broken i.e. 19.5.1 and 19.5.2 = Black screen of death. Drivers from Jan to April 2019 are okay. I may try DDU (Display Driver Uninstaller) before installing 19.5.2.

----

Meanwhile, my Surface Pro with Intel HD 640 Iris Pro drivers are reasonably stable e.g. no blue screen of death with playing youtube videos.

RX-5700 is slightly faster than Vega 64 on AMD game title, just as I expected when Vega 56 was over-clocked to 1710 Mhz.

RX-5700 is half the chip area size from RTX 2070's 445 mm2, hence RX-5700 has 225 mm2 which close to RX-480's 232 mm2 and 7870's 210 mm2 which are PS4/PS4 Pro GPU chip area sizes.

That sucks. Got my old relatives a Ryzen 3 $ 270 laptop with 8GB RAM a year ago. Hope it doesn't conk out on them. It's just for web browsing and email checking.

Have an old Surface Pro and the problem I have is the usb port is dying and the charge port or charger is almost broken. I really wish the next surfaces will have 2 usb ports for data and charging on both.

I can repair or replace hardware, but I can not repair close source driver issues. I have repaired Surface Pro's USB ports. Repairing PC hardware is not my day job i.e. just a hobby from my soldering my Amiga PCB days.

Google Chrome with AMD RX Vega mobile driver + Youtube in 17.x version OEM driver range has higher BSOD frequency. Buggy video decode accelerator drivers.

This Ryzen mobile RX Vega IGP issue has blown up in social and mainstream news media sources last year.

------

Furthermore, AMD has abandoned Enduro switchable graphics with 19.5.1 driver being the last driver.

https://www.neowin.net/news/amd-radeon-1951-driver-delivers-support-for-windows-10-may-2019-update-and-rage-2/

Meanwhile, this release marks the end of AMD's support for its Mantle API and Enduro switchable graphics technology.

If there's a requirement for discrete mobile GPU, I wouldn't buy AMD. I plan to use AMD to lower the cost for Intel base laptops, but I'm not planning to buy mobile AMD laptops.

AMD Ryzen CPUs are fine, but mobile RX Vega IGPs are problematic.

@ronvalencia: Oh wow. I had a PC with a Ryzen 5 2400G that would blue screen on YouTube videos often. I didn’t keep that chip very long but never really knew there was a driver issue.

My issue is for fleet deployment and I was testing a HP laptop with Ryzen mobile laptop for stability. My conclusion, not good enough for enterprise deployment. I don't care about Linux on the client side.

Personally, I have saved up money for my next 2-in-1 laptop.

I'm looking forward with Intel Ice-Lake 2-in-1 tablet/laptops. I wished NVIDIA has X86 CPU i.e. buy out VIA X86 CPU business.

With game consoles, we have Microsoft and Sony doing QA testing/modifications on AMD GPU hardware and software stack.

8 core without SMT at 2.8-3.2GHz all core boost... Is my best guess as Zen mobile only offers 4 core chips at the moment so two 4 core chips on one die without the SMT to lower TDP and increase the core clock.

Them going with 8 cores and 16 threads if it even means a 200MHz sacrifice would be a HUGE mistake since sub 3GHz CAN be a bottleneck even at 1440/60Hz, disable SMT and push core clocks if these guys are serious about giving us a real gaming console.

Didn't even know AMD had switchable graphics. Was it for systems with both AMD iGPUs and dGPUs?

Even if there won't be a zen mobile 8 core in 2020. Sony can always request for a custom part just as how there were no 8 core jaguars for mainstream mobile even after a few years after PS4's release.

I am actually interested on which would be better for coded to metal gaming and at which clockspeeds should they meet at to perform roughly similarly. Would a 16c/16t 2.8 ghz beat an 8 core 3.2 ghz? It would be an interesting case. I believe this is hard to really calculate right now as games are currently designed to baseline 6 cores / "6.5 cores?" with no SMT just as how I don't think a GTX 2080Ti isn't fully used to meter its power by only using sharper shadows/higher resolution of current gen games.

A reason I can think of consoles going for 8c/16t mobile ryzen would be for marketing that 16 thread might look better to the average joe specially if console makers don't advertise CPU caches/clockspeeds on the box.

https://www.xbox.com/en-US/xbox-one-x

https://www.playstation.com/en-nz/explore/ps4/tech-specs/

Then again that might be less relevant as the target group might be finally into to online info. Don't know if I'm overestimating the general target though.

Part of me thinks that if we ever get a full desktop ryzen with the same cache as the desktop equivalent and threads, the clockspeeds would be really gimped.

For now, I'm leaning towards similar high 2s-low 3s but for custom Ryzen mobile of 8 cores/8 threads and gimped cache compared to desktop.

It'd be interesting to see these compete with through next gen Sony vs Xbox:

One uses a 10 core, 10 thread 3.2 Ghz, half l3 cache, vs 8 core, 16 thread, 2.8 Ghz with full l3 cache.

Unlikely but I'd like to see that compete on closed systems.

I'm not sure if it has to do with threads but wasn't amd beating intel due to sheer threads for hosting a stream? I do wonder if Sony and MS will have that as part of the next gen agenda, performance while streaming and far more sophisticated built in editing into the systems thanks to our current Youtube/Twitch age. Super extrapolated from the Share button.

Ryzen 2 with some 3 features mix in 8 cores 16 threads who knows what speed.

I'll guess nothing below X's 2.3 Ghz, mostly for marketing purpose.

RTX 2080 Ti is not fully used since DirectML is needed to enable rapid pack maths accessible for both recent AMD and NVIDIA GPUs.

"World War Z" game example has used AMD's Rapid Pack Path feature on RX Vega GPUs via AMD's extensions in Vulkan API.

• Above slide from NVIDIA's presentation for Turing GPUs.
• DirectML is better than hand written shaders.
• DirectML's Metacommands are similar to direct hardware access from AMD's Shader Intrinsic Functions and NVIDIA's NVAPI Intrinsic Functions.
• DirectML is also maps to NVIDIA's massive TFLOPS Tensor cores to addition to rapid pack maths features on Turing CUDA cores.
• All NVIDIA Turing GPUs has rapid pack maths features on Turing CUDA cores as baseline specifications.

Metacommands completes DirectX12's features for near future Windows Gaming PCs to be similar to Xbox One in regards to low level graphics API access.

AMD's Lisa Su confirms both Xbox Anaconda and future XCloud has deep learning rapid pack maths feature.

There's a reason on why NVIDIA has included rapid pack maths path in all Turing GPUs since both NVIDIA and AMD common road maps from Microsoft.

---------

With Zen v2's hyper-threading...

• Background tasks can be done on secondary weaker threads while the games has primarily threads, hence future game consoles may have full 8 CPU core access.
• May eliminate ARM CPU to reduce IP payment cost. PS4 has 8 Jaguar CPUs (i.e. 6.5 CPUs for games, 1.5 CPUs for OS) and ARM Cortex CPU (for background downloading/internet related tasks), hence PS4 technically has at-least 9 CPU threads. 8 core Zen v2 has 16 threads.
• AMD doesn't gimp hyper-threading features like Intel. Ignore Intel's product segment BS. Zen v2 has two extra SKUs above 8 core/16 threads CPUs which are 12 core/24 threads and 16 core/32 threads Zen v2 SKUs for socket AM4.

---------

I have Intel IGP and AMD GCN switchable laptop.

Been out of the PC game for awhile I see!

@ronvalencia said:

@rzxv04 said:

Didn't even know AMD had switchable graphics. Was it for systems with both AMD iGPUs and dGPUs?

@Grey_Eyed_Elf said:

8 core without SMT at 2.8-3.2GHz all core boost... Is my best guess as Zen mobile only offers 4 core chips at the moment so two 4 core chips on one die without the SMT to lower TDP and increase the core clock.

Them going with 8 cores and 16 threads if it even means a 200MHz sacrifice would be a HUGE mistake since sub 3GHz CAN be a bottleneck even at 1440/60Hz, disable SMT and push core clocks if these guys are serious about giving us a real gaming console.

Even if there won't be a zen mobile 8 core in 2020. Sony can always request for a custom part just as how there were no 8 core jaguars for mainstream mobile even after a few years after PS4's release.

I am actually interested on which would be better for coded to metal gaming and at which clockspeeds should they meet at to perform roughly similarly. Would a 16c/16t 2.8 ghz beat an 8 core 3.2 ghz? It would be an interesting case. I believe this is hard to really calculate right now as games are currently designed to baseline 6 cores / "6.5 cores?" with no SMT just as how I don't think a GTX 2080Ti isn't fully used to meter its power by only using sharper shadows/higher resolution of current gen games.

A reason I can think of consoles going for 8c/16t mobile ryzen would be for marketing that 16 thread might look better to the average joe specially if console makers don't advertise CPU caches/clockspeeds on the box.

https://www.xbox.com/en-US/xbox-one-x

https://www.playstation.com/en-nz/explore/ps4/tech-specs/

Then again that might be less relevant as the target group might be finally into to online info. Don't know if I'm overestimating the general target though.

Part of me thinks that if we ever get a full desktop ryzen with the same cache as the desktop equivalent and threads, the clockspeeds would be really gimped.

For now, I'm leaning towards similar high 2s-low 3s but for custom Ryzen mobile of 8 cores/8 threads and gimped cache compared to desktop.

It'd be interesting to see these compete with through next gen Sony vs Xbox:

One uses a 10 core, 10 thread 3.2 Ghz, half l3 cache, vs 8 core, 16 thread, 2.8 Ghz with full l3 cache.

Unlikely but I'd like to see that compete on closed systems.

I'm not sure if it has to do with threads but wasn't amd beating intel due to sheer threads for hosting a stream? I do wonder if Sony and MS will have that as part of the next gen agenda, performance while streaming and far more sophisticated built in editing into the systems thanks to our current Youtube/Twitch age. Super extrapolated from the Share button.

RTX 2080 Ti is not fully used since DirectML is needed to enable rapid pack maths accessible for both recent AMD and NVIDIA GPUs.

"World War Z" game example has used AMD's Rapid Pack Path feature on RX Vega GPUs via AMD's extensions in Vulkan API.

• Above slide from NVIDIA's presentation for Turing GPUs.
• DirectML is better than hand written shaders.
• DirectML's Metacommands are similar to direct hardware access from AMD's Shader Intrinsic Functions and NVIDIA's NVAPI Intrinsic Functions.
• DirectML is also maps to NVIDIA's massive TFLOPS Tensor cores to addition to rapid pack maths features on Turing CUDA cores.
• All NVIDIA Turing GPUs has rapid pack maths features on Turing CUDA cores as baseline specifications.

Loading Video...

Metacommands completes DirectX12's features for near future Windows Gaming PCs to be similar to Xbox One in regards to low level graphics API access.

AMD's Lisa Su confirms both Xbox Anaconda and future XCloud has deep learning rapid pack maths feature.

There's a reason on why NVIDIA has included rapid pack maths path in all Turing GPUs since both NVIDIA and AMD common road maps from Microsoft.

---------

With Zen v2's hyper-threading...

• Background tasks can be done on secondary weaker threads while the games has primarily threads, hence future game consoles may have full 8 CPU core access.
• May eliminate ARM CPU to reduce IP payment cost. PS4 has 8 Jaguar CPUs (i.e. 6.5 CPUs for games, 1.5 CPUs for OS) and ARM Cortex CPU (for background downloading/internet related tasks), hence PS4 technically has at-least 9 CPU threads. 8 core Zen v2 has 16 threads.
• AMD doesn't gimp hyper-threading features like Intel. Ignore Intel's product segment BS. Zen v2 has two extra SKUs above 8 core/16 threads CPUs which are 12 core/24 threads and 16 core/32 threads Zen v2 SKUs for socket AM4.

---------

I have Intel IGP and AMD GCN switchable laptop.

Thanks.

Sadly, I don't understand most of it.

I guess APIs are now better for PC but I was talking more about game designs leveraged only when AAA devs start building exclusively for 9th gen. It gets a little murky with how you can technically run 9th gen exclusives on much weaker pc and sacrifice visuals and performance but are technically feasible.

Does the Xbox also have ARM cpu?

Not sure if it's technically gimping but I only see it as a product segmentation like how most Ryzen 3 desktop don't have SMT. AMD probably sees it too potent for the price at least for now.

Rapid Pack Math feature enables recent GPU designs to execute 16 bit floating point/integer datatypes at twice the throughput rate over 32bit floating point.

Newer GPUs like Vega II and Turing GPUs has the ability execute 8 bit integer datatypes at four times the throughput rate over 32bit floating point, these are usually associated to AI/deep learning workloads.

Older DirectX12 GPUs execute DirectML workloads with no throughput upgrades.

-------------

Xbox One's APU has ARM based security processor.

Taken from AMD site:

AMD gives you a dedicated platform security processor (PSP) built into your AMD APU. The ARM TrustZone layer runs on top of the hardware, creating a secure environment by partitioning the CPU into two virtual ”worlds.” Sensitive tasks are run on the PSP – in the "secure world" – while other tasks are run in the "typical operation." This helps ensure the secure storage and processing of sensitive data and trusted applications. It also helps protect the integrity and confidentiality of key resources, such as the user interface and ​service provider assets.

PS4 has an extra ARM CPU for internet background workloads besides AMD's platform security processor (ARM Cortex A5 based).

It's unknown if AMD will use it's K12 ARM clone to replace ARM Ltd's ARM CPU IP. AMD's K12 ARMv8 64bit clone R&D recycled Ryzen's RISC core design with ARM instruction set instead of X86 instruction set.

Intel disabled hyper-threading on 6 cores i5-9600K and 8 cores i7-9700K ($399) which are stupid since AMD's Ryzen 7-3700 and 7-3800 ($399) has 16 threads.

https://www.rockpapershotgun.com/2019/05/28/amd-ryzen-3000-release-date-price-specs/

I received a full store credit for my warranty returned MSI x299 Carbon AC motherboard which is more than enough for buying AM4 X570 motherboard.

I only need to pay for Zen v2 CPU and I already have ARGB DDR4-3600 memory modules. I'm aiming for 16 core Zen v2.

--------

ARGB = Addressable Red Green Blue LEDs.

Thanks. So Rapid Packed Math is also in Turing and is the way moving forward?

Thanks. So basically PS4 has an actual separate physical ARM cpu while Xbox has it baked into the APU itself.

Rapid Packed Math feature also exist on baseline Turing GPU's CUDA cores besides Tensor cores.

Both PS4 and XBO APUs has AMD's PSP.

Sony added extra ARM SoC for internet related background workloads.

Both PS4 and XBO APUs has AMD True-Audio DSPs which are 3rd party IP licence from Cadence Design Systems.

From Polaris GCN, AMD added extra DSP instructions to execute True-Audio on it's GPUs and re-market it as True-Audio Next (which includes Radeon Rays 3.0 for ray-tracing audio, runs better with Rapid Pack Maths on Vega GCN), hence reducing the extra IP payments.

Thanks. I guess PS4 has more CPU package then but I guess it's clocked slower.

I wonder if the ip license is for x-amount of time rather than per item sold.

In my experience, PS4 base, the OS is much slower, specially in screenshots/recording than base Xbox One (not even S). I guess MS just has better software outside games.