|
GRINDCORE MEGGIDO posted:How different is context switching in console os's Vs Windows? Not especially, user facing stuff like games are run at a lower privilege level, for security and anti-piracy reasons. The XBox/PS4 are basically fancy mATX PCs running some proprietary blend of either FreeBSD or Windows under the hood anyways, so it's however the base kernel they decided to use works.
|
#
?
Jan 9, 2019 00:05
|
|
|
|
| # ? Apr 26, 2024 01:12 |
|
|
Crotch Fruit posted:Consider that a GTX 1080 has 2560 CUDA cores and 160 texture units Also the cores on the GPU are more like "cores" in a marketing sense rather than a realistic and practical one in comparison to the CPU when talking about flexibility and capability. Laziness also isn't the issue when talking about programming either. Most of those guys are getting worked pretty hard, to the point of premature burn out sometimes. Its almost always either comes down to development budget or market support when talking about taking advantage of hardware features like more CPU cores or even other stuff like more system RAM and SSD's.
|
|
#
?
Jan 9, 2019 01:32
|
|
|
Parallelism is hard, but it doesn't have to be. In the Rust programming language there's a massive amount of work being done to make parallelism easier to implement, this work spans from language design to third party libraries like Rayon and entire game engines like Amethyst that are designed to let anyone have access to parallelism. I would expect these efforts to be mirrored for the latest C++ spec, but I don't have any involvement or understanding of the C++ ecosystem. The tools now exist, it's just a matter of someone adopting them.
|
|
#
?
Jan 9, 2019 05:59
|
|
|
Measly Twerp posted:Parallelism is hard, but it doesn't have to be. In the Rust programming language there's a massive amount of work being done to make parallelism easier to implement, this work spans from language design to third party libraries like Rayon and entire game engines like Amethyst that are designed to let anyone have access to parallelism. Usually the writing part is easier than figuring what can be parallelized in the first place.
|
|
#
?
Jan 9, 2019 06:47
|
|
|
PC LOAD LETTER posted:Laziness also isn't the issue when talking about programming either. Most of those guys are getting worked pretty hard, to the point of premature burn out sometimes. Its almost always either comes down to development budget or market support when talking about taking advantage of hardware features like more CPU cores or even other stuff like more system RAM and SSD's. Yeah quad cores were the norm for a full decade and taking advantage of a lot of CPU cores is difficult, why put in the effort for something that very few people have? Only in the past couple years has this made sense and we have seen some change.
|
|
#
?
Jan 9, 2019 07:12
|
|
|
A lot of easy(ish) parallelism can be had using a broker/worker system which a lot of game engines are moving towards. Break the work into small chunks with the broker thread(s), then farm the little tasks out to worker threads who do the work and hand it back over. Issue is in a lot of cases the chunking is non-trivial to do, so outside of the specific game framework it's designed for, it quickly stops working at all. Even simple things like path finding can become hard to multi-thread when you start having to deal with inter-dependent paths (where entities can collide or obstruct each other) or pathing that's costly enough to span several game ticks per thread worker.
|
|
#
?
Jan 9, 2019 09:53
|
|
|
e: Never mind this was confusing and dumb speculation about a derail and right before AMD has a keynote. Hoping for some specs today, will be happy with just a release target.
|
|
#
?
Jan 9, 2019 15:49
|
|
|
|
|
#
?
Jan 9, 2019 16:12
|
|
|
I'm ready for disappointment! At least Raja won't be there to talk about some bollywood bullshit and shrouds for 2 hours
|
|
#
?
Jan 9, 2019 16:16
|
|
|
https://www.youtube.com/watch?v=bibZyMjY2K4
Risky Bisquick fucked around with this message at 18:05 on Jan 9, 2019 |
|
#
?
Jan 9, 2019 16:22
|
|
|
Risky Bisquick posted:https://www.youtube.com/watch?v=bibZyMjY2K4 We're live 
|
|
#
?
Jan 9, 2019 18:06
|
|
|
this isn't lisa su 
|
|
#
?
Jan 9, 2019 18:08
|
|
|
SHE IS WEARING A BLACK LEATHER JACKET LOL
|
|
#
?
Jan 9, 2019 18:12
|
|
|
eames posted:SHE IS WEARING A BLACK LEATHER JACKET LOL COMING FOR THE CROWN 
|
|
#
?
Jan 9, 2019 18:14
|
|
|
Come on 64C/128T 3990WX!
|
|
#
?
Jan 9, 2019 18:15
|
|
|
I JUST LEARNED WHAT A ZETTABYTE IS
|
|
#
?
Jan 9, 2019 18:16
|
|
|
This is a lot of talking, toss the water bottle now!!! https://www.youtube.com/watch?v=CH-v0BfEf9I
|
|
#
?
Jan 9, 2019 18:17
|
|
|
NewFatMike posted:COMING FOR THE CROWN You know I had to Su it to 'em
|
|
#
?
Jan 9, 2019 18:28
|
|
|
zzzz
|
|
#
?
Jan 9, 2019 18:29
|
|
|
Well now I know how to pronounce Huawei. Learned nothing new about AMD products though.
|
|
#
?
Jan 9, 2019 18:30
|
|
|
Show us Zen 2 already goddamn it! 
|
|
#
?
Jan 9, 2019 18:32
|
|
|
CES -> Chromebooks using A4/A6 
|
|
#
?
Jan 9, 2019 18:34
|
|
|
EoRaptor posted:Well now I know how to pronounce Huawei. Learned nothing new about AMD products though. I missed it, but I have always pronounced it "wah-way." Is this incorrect?
|
|
#
?
Jan 9, 2019 18:36
|
|
|
hoo aaah way, the first two sounds are the same timeframe as saying the second
|
|
#
?
Jan 9, 2019 18:36
|
|
|
I guess "Creator" is the new "Prosumer"... 
|
|
#
?
Jan 9, 2019 18:37
|
|
|
You've disappointed me Lisa, you've brought me no zen 2 (yet), but a man of wrong ram.
|
|
#
?
Jan 9, 2019 18:39
|
|
|
She's announcing the RX 590 again
|
|
#
?
Jan 9, 2019 18:42
|
|
|
I'm guessing they're saving Zen 2 until the very end of the presentation.
|
|
#
?
Jan 9, 2019 18:43
|
|
|
ratbert90 posted:I missed it, but I have always pronounced it "wah-way." Is this incorrect? Nope, that's about right. I was mostly being facetious, I've heard it pronounced incorrectly a bunch of times, and nothing about AMD's future was actually being said.
|
|
#
?
Jan 9, 2019 18:45
|
|
|
Radeon 7 announced blah GPU e: 60 CU 1.8 ghz vega2, same power as vega. 16gb hbm2
|
|
#
?
Jan 9, 2019 18:45
|
|
|
Risky Bisquick posted:Radeon 7 announced blah GPU 16 GB HBM 
|
|
#
?
Jan 9, 2019 18:46
|
|
|
Vega on 7nm 
|
|
#
?
Jan 9, 2019 18:46
|
|
|
Vega 64 at +35% performance is still bad, right I can't remember seeing Vega 64 benchmarks
|
|
#
?
Jan 9, 2019 18:48
|
|
|
ITS THE CREAM OF THE CROP
|
|
#
?
Jan 9, 2019 18:48
|
|
|
It's a 1080ti with more ram
|
|
#
?
Jan 9, 2019 18:48
|
|
|
DX12, Vulkan and GPU Compute increases... looks like more FineWine  Still looking forward to the Zen 2 reveal. Edit: On par with a 2080... not bad but the price will be the big determining factor. Icept fucked around with this message at 18:51 on Jan 9, 2019 |
|
#
?
Jan 9, 2019 18:49
|
|
|
.
sincx fucked around with this message at 05:50 on Mar 23, 2021 |
|
#
?
Jan 9, 2019 18:49
|
|
|
amd stock dropped 5% following the R7 reveal
|
|
#
?
Jan 9, 2019 18:50
|
|
|
Risky Bisquick posted:It's a 1080ti with more ram which would be fine at $400 but it feels like the bulk cost of the RAM is more than that vvv AMD recommends RTX on
eames fucked around with this message at 18:53 on Jan 9, 2019 |
|
#
?
Jan 9, 2019 18:50
|
|
|
|
| # ? Apr 26, 2024 01:12 |
|
|
They were comparing it to a 2080, so in real benchmarks it will only be 25% slower.
|
|
#
?
Jan 9, 2019 18:51
|
|