It stands to reason that the driver processes will be assigned to underutilized cores. Most of these benchmarks are showing the B580 bogging down on older 6-8 core CPUs.
It stands to reason that the driver processes will be assigned to underutilized cores. Most of these benchmarks are showing the B580 bogging down on older 6-8 core CPUs.Sounds like a question for @DAPUNISHER . My guess would be it is more of an IPC than a thread thing but I have no idea.
It stands to reason that the driver processes will be assigned to underutilized cores. Most of these benchmarks are showing the B580 bogging down on older 6-8 core CPUs.
stop saying that! they are not old!!bogging down on older 6-8 core CPUs.
stop saying that! they are not old!!
I have a 5600 and a 5700x. The 5600 should be oc'd to 4.6-4.7ghz like the 5600x. The only difference is a little smoother performance and a few fps better with 8 cores. I thought 6core 12 threads would be worse performance. On Zen 3, it does not make a significant difference. On a new build I would say go with an 8 core CPU, but 6 core CPU's are still very capable.Do we have any data on how B580 performs with extra cores? For example, 5900X vs 5600X?
Yes I will not be surprised they can use SW Tricks with their HWAre you wondering if there is some secret sauce with all team blue test systems?
stop saying that! they are not old!!
That doesn't really address the problem, though. The B580 is fighting for CPU resources and can't perform well when it can't get them. Adding moar coars should solve the problem, no? For example, bench B580 on 12900k vs 5800X3D or 7800X3D (or similar). X3D CPUs can do quite well against the 12900k with a high-end dGPU, but with B580 are we going to see a different picture?I have a 5600 and a 5700x. The 5600 should be oc'd to 4.6-4.7ghz like the 5600x. The only difference is a little smoother performance and a few fps better with 8 cores. I thought 6core 12 threads would be worse performance. On Zen 3, it does not make a significant difference. On a new build I would say go with an 8 core CPU, but 6 core CPU's are still very capable.
A lot of these game testers do things as a gimmick for views without taking a big picture look. I do not see the benefit of GPU reviews where 1080P is still the backbone of the review in 2024/2025. 1440P and 4K are more important but look less impressive or different in GPU reviews.
Gaming at 1440P even on low settings is far superior to 1080P at any level of detail. I also think that med/high settings in reviews should be the focal points because that represents real world settings by most gamers. Going with ultra settings is good if you have 12-16GB+ VRAM.
That's interesting and would make good content. However, It does not resolve the issues the B580 is facing as an entry level card. PCIe 3, rebar, older CPUs, all rob performance disproportionately v. the competition. Upscaling and ray tracing are advantages over AMD, that vanish without a beefy enough CPU.That doesn't really address the problem, though. The B580 is fighting for CPU resources and can't perform well when it can't get them. Adding moar coars should solve the problem, no? For example, bench B580 on 12900k vs 5800X3D or 7800X3D (or similar). X3D CPUs can do quite well against the 12900k with a high-end dGPU, but with B580 are we going to see a different picture?
Apparently Intel IS dropping a faster part, if the rumor I found is to be believed. It will cost $500-$600 and performance will be where you expect for a part in that range. Unconfirmed source, but given they just announced a launch event, it makes sense.
Wish I had made more effort to follow them. Might’ve been able to dig up juicier stuff.
It's pretty cheap, though, and can also be roughly approximated with a 13700k for anyone feeling courageous enough to run one of those things. Don't buy em used.The 12900K is only 3yrs old.
Can you share the link for the source?Apparently Intel IS dropping a faster part, if the rumor I found is to be believed. It will cost $500-$600 and performance will be where you expect for a part in that range. Unconfirmed source, but given they just announced a launch event, it makes sense.
Wish I had made more effort to follow them. Might’ve been able to dig up juicier stuff.
By Zen 6 it will definitely be obsolete in performance thougheh in CPU years 4 years (Zen 3) is on the older side considering there have been two generations since. Still very happy with mine.
Not a big issue. He can get a boost with a cheap 5700X3D without changing whole platform.By Zen 6 it will definitely be obsolete in performance though
Who changed the dictionary and didn't tell me?By Zen 6 it will definitely be obsolete in performance though
If the 265k is cheap enough that would be another option. Right now it's $360 (approximately) so the 12900k has it beat on price.12900K is the most risk free option or a 265K if you believe Intel's BIOS Update
If the driver is so reliant on the CPU, and especially the uarch performance, then Intel solutions with E cores could potentially be catastrophic, the opposite of what people are expecting. You would be way behind Zen 2 levels.If the 265k is cheap enough that would be another option. Right now it's $360
Regardless the idea would be to get some data on whether the B580 can benefit from a multicore CPU with more than 6-8 cores. In fact, just benching a 12900k with e-cores enabled and disabled with B580 and some roughly-equivalent NV/AMD video cards, would be illuminating.
Yes I just want to be sure they are not doing shenanigansIf the driver is so reliant on the CPU, and especially the uarch performance, then Intel solutions with E cores could potentially be catastrophic, the opposite of what people are expecting. You would be way behind Zen 2 levels.
Instead I would assume disabling the E cores would make it perform quite a bit better.
There's no evidence to suggest that it would perform dramatically better on Intel platforms, and this is going to feel bad for some people - other than hopium.
Not a big issue. He can get a boost with a cheap 5700X3D without changing whole platform.
If the driver is so reliant on the CPU, and especially the uarch performance, then Intel solutions with E cores could potentially be catastrophic, the opposite of what people are expecting. You would be way behind Zen 2 levels.
Instead I would assume disabling the E cores would make it perform quite a bit better.
There's no evidence to suggest that it would perform dramatically better on Intel platforms, and this is going to feel bad for some people - other than hopium.
Looked at the important parts.HWUB retests in a variety of conditions with 5600 vs 9800X3d:
Some of these architectural decisions may be due to their crazy focus on trying to save on costs/die size and taking shortcuts. Ironically, because their hardware sucks, it ended up being bloated anyways.Very well written! Hopefully, these insights are allowed past the marketing department, managers, and other bureaucratic mechanisms that "protect" the developers. Hopefully, the driver team has enough bandwidth to address this, compared to working on new features, new hardware, or fixing other bugs.
Some of the issues you can see in the graphs indicate architectural issues, such as batching strategy and DMA chunk sizes. Other things could be hardware choices, such implementing less of the DMA mastering on the GPU side (i.e., using Windows paging to transfer data to the GPU instead of sending a physical address to the GPU DMA engine).
The above comment implies the overall driver code is terrible.That ISR/DPC graph is crazy. Spending 1/3 of a CPU-second in ISRs for every 1 second of wall clock time running a single threaded DX11 application is insane. Even if they moved that into a DPC, that's still going to be disruptive to other latency-sensitive applications on the system. I'd be curious to know how evenly distributed they are across CPUs and whether it's a lot of short ISRs or a few long ones that add up.
It's unlikely that it's contributing significantly to the poor 3DMark API Overhead test results at these kind of frame rates, but it's certainly interesting that Intel can't seem to do Independent Flip on Vulkan swapchains. Having to context switch to DWM to present each frame isn't free and almost certainly is costing them significant performance in very high frame rate Vulkan applications. Nvidia's driver can also directly present Vulkan swapchains like you saw on the AMD test.