Scheduler improvements. These improvements will be standard in an upcoming stable release, but with the launch of zen 5, amd worked with microsoft to get an optional patch released. Wendell with level1techs has some good content on the subject if you want to learn more.
The X3D’s just have ridiculously overbuilt caches. That’s the main thing (that I’m aware of) that makes them so fast with games while still being cheap.
AMD just put more of that than Windows could handle to max out speed in games.
So I’m assuming (and may be wrong) that the windows optimization is just updating it so that it could plan for those insane caches and take full advantage of it.
That would make sense! I’m just glad they’re not doing their usual thing of not properly testing, not working with Microsoft, and then blaming everyone else when something breaks
I mean. I’m just assuming because linked article doesn’t give details and I’m not trying to hunt it down.
But that’s a likely reason why we could see 10% gains for the best gaming chip currently out from just a software update
Like, 10% doesn’t sound like much to people, but gamers will traditionally drop a couple hundred bucks for that type of gain, to get it on literally the best chip tho would likely be closer 4-500 because the price of performance gains is usually exponential at that level
Yeah. I actually read a good article last night but didn’t save it.
7800x3d beats the 9k next gen chips in games, because the 7800x3d has one “group” of cores.
Next gen has basically double cores, but that means two “groups”. So while for production work, yeah, it goes 2x as fast because it has 2x the cores.
But for games and fps, latency is the big deal.
So spreading between two core groups hurts by making additional latency for the two groups to talk to each other.
Makes me think they’ll be another update to disable one set to remove that latency, or something to split the workload so the groups work on different things and that latency ain’t an issue. Another software fix to get them the same game performance as 7800x3d shouldn’t be hard, would be like how back in the day some chips would disable cores for games because games didn’t use them. They just got to flip off one set of cores instead of individual cores.
Like, the new 9ks are like a work truck, it ain’t fast, but if you’re moving a truck is what you need to move lots of stuff quickly.
7800x3d is like a Honda Civic. Less HP but way less weight and hauling capacity.
Production work needs that hauling capacity, but games is more like running across town with no cargo, it just has to make the trip in the shortest amount of time
I wonder what that fix is tho
Turns out AMD got off it’s ass and started working together with microsoft instead of complaining
nice
Scheduler improvements. These improvements will be standard in an upcoming stable release, but with the launch of zen 5, amd worked with microsoft to get an optional patch released. Wendell with level1techs has some good content on the subject if you want to learn more.
ey nice! Thankz! I’ll have a look at that YouTube channel!
The X3D’s just have ridiculously overbuilt caches. That’s the main thing (that I’m aware of) that makes them so fast with games while still being cheap.
AMD just put more of that than Windows could handle to max out speed in games.
So I’m assuming (and may be wrong) that the windows optimization is just updating it so that it could plan for those insane caches and take full advantage of it.
That would make sense! I’m just glad they’re not doing their usual thing of not properly testing, not working with Microsoft, and then blaming everyone else when something breaks
I mean. I’m just assuming because linked article doesn’t give details and I’m not trying to hunt it down.
But that’s a likely reason why we could see 10% gains for the best gaming chip currently out from just a software update
Like, 10% doesn’t sound like much to people, but gamers will traditionally drop a couple hundred bucks for that type of gain, to get it on literally the best chip tho would likely be closer 4-500 because the price of performance gains is usually exponential at that level
The whole X3d story is just crazy
Oh I know, I’d spend that for 1%
Yeah. I actually read a good article last night but didn’t save it.
7800x3d beats the 9k next gen chips in games, because the 7800x3d has one “group” of cores.
Next gen has basically double cores, but that means two “groups”. So while for production work, yeah, it goes 2x as fast because it has 2x the cores.
But for games and fps, latency is the big deal.
So spreading between two core groups hurts by making additional latency for the two groups to talk to each other.
Makes me think they’ll be another update to disable one set to remove that latency, or something to split the workload so the groups work on different things and that latency ain’t an issue. Another software fix to get them the same game performance as 7800x3d shouldn’t be hard, would be like how back in the day some chips would disable cores for games because games didn’t use them. They just got to flip off one set of cores instead of individual cores.
Like, the new 9ks are like a work truck, it ain’t fast, but if you’re moving a truck is what you need to move lots of stuff quickly.
7800x3d is like a Honda Civic. Less HP but way less weight and hauling capacity.
Production work needs that hauling capacity, but games is more like running across town with no cargo, it just has to make the trip in the shortest amount of time
Shits pretty interesting.
I guess those optimizations might be similar to the integration in windows 11 for the Intel big/little thing