[Help] Windows 10 Woes and issue (still)

DX9? 10 years ago!

 

It is but to a much lesser degree. I just felt this might be worth bringing up.

 

I'm curious: do you really need that many cores? Cores only help if you are going to either run a lot of things, or if your programs can do multiprocessing to that degree.
On my end, I think I would have attempted to get a better GPU (than what I have now) if I ever prioritized gaming over my work.

As @Bad_Ad84 wrote, there would have been a lot more complains if the drivers were indeed bad. So this has to be an isolated case.

Have you tried doing anything about your RAM yet? I am out of date (it used to still be FSB, when I meddled with OC), but the RAM used to also affect things like your system bus.

 

I do actually make use of them. Incidentally the only gaming aspect I have that utilizes them for gaming is RPCS3 as it favors core count.

I did run memtest for 1 pass and it returned no errors. I may take the sticks out soon and try swapping em around (or removing a pair altogether)

 

1 pass isn't sufficient.

 

I'll run it for 7 then. That should be sufficient?

 

Depends how often the fault occurs. I've had sticks that only fail 1 in 10 passes.

But certainly run it at least 5. Ideally much longer

 

Almost 16 years, actually. DX9 was first released in 2002.

 

Ill run it while im sleeping then. I tend to knock out for 8+ hours and it seemed to get 1 pass per hour

Noted for a fun idea in the future

 

Saw this and it's kind of related.

https://www.reddit.com/r/Games/comments/7nqgpb/intel_cpus_processor_design_flaw_may_cause_5_to/

As mentioned here in this article:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

As quoted:

A fundamental design flaw in Intel's processor chips has forced a significant redesign of the Linux and Windows kernels to defang the chip-level security bug.

Programmers are scrambling to overhaul the open-source Linux kernel's virtual memory system. Meanwhile, Microsoft is expected to publicly introduce the necessary changes to its Windows operating system in an upcoming Patch Tuesday: these changes were seeded to beta testers running fast-ring Windows Insider builds in November and December.

Crucially, these updates to both Linux and Windows will incur a performance hit on Intel products. The effects are still being benchmarked, however we're looking at a ballpark figure of five to 30 per cent slow down, depending on the task and the processor model. More recent Intel chips have features – specifically, PCID – to reduce the performance hit.

So, for those who enjoyed playing video games on PCs with Intel-based CPUs, be sure to note that some of you may be one of those users who will expect to experience performance drops. Of to what extent, I don't know, your mileage may vary vastly.

Since this is related to video games, high performances on PCs, and CPUs in general, is this allowed?

 

I researched this a little, and one source said that Haswell and newer processors have this. However, I looked at the CPU info on my PC - which has an Ivy Bridge processor - and it says it does have PCID. I checked on my MacBook Pro as well, which has a Haswell CPU, and it also says it has PCID. So, probably if your processor is an Ivy Bridge or newer, it has PCID. Not sure about Sandy Bridge, because I don't have any Sandy Bridge machines.

From what I've read, it appears the security flaw doesn't affect AMD processors, but the fixes that companies are planning on releasing may still have an effect on AMD systems.

Also, Apple supposedly has already partially fixed this with the macOS 10.13.2 update, which was released in early December. I've had this update for a while and haven't seen any noticeable slowdown.

 

Ran for 6 passes. No errors but it got a lot slower as it ran.

I had read this as well and it makes me wonder why if it's unrelated to the issue.

 

It's the rational fear that the rightfully-unaffected minority would have to bear the brunt of the consequences to the problem suffered by the majority (e.g. because it is easier to design a slower kernel that works for everybody, rather than making multiple versions). But as Windows has not received its share of fixes yet, it is too early to say that it will surely happen.

Logic says it can be avoided though.

From what I understand, the "performance dip" is also very subjective (to what the program does) and may concern just privileged code execution. So there may be little to no dip in performance.

 

AMD submitted a patch to the linux kernel to disable the fix on AMD processors.

https://lkml.org/lkml/2017/12/27/2


But apparently AMD processors DO have the bug. But it only leaks data to the same process. Not cross process like the intel one (imagine being able to steal data from a different VM in cloud computing, etc).

 

Wow that is REALLY bad

 

There are 2 different security flaws. One is called Meltdown, and affects only Intel processors. The other one is called Spectre, which affects Intel, AMD, and ARM processors - so, pretty much everything.

 

Yes, Meltdown that accesses cross processes. (Intel only)
Spectre that is only for that process (all, but can be solved in software).

Also apparently the google proof of concepts for spectre only worked on all 3 (intel, amd and arm) with the least malice. The nasty one only worked on Intel too.

 

Flaws and bugs they call them, I say it was by design. =)

 

9 months, 20+ emails, 12 different representatives taking over, about a month of shipping time (9 of which are scheduled delivery days taking 3 days each because it was going to arrive before those dates), and multiple angry emails about the 20+ emails not being addressed properly (only 4 of them were read), we arrive with this small excerpt.


Whether this is truly the solution or not I dont know.

And in other news, some users have experienced some "bricking" issues with Ryzen based systems with MS deployed patches for the various issues that showed up over the past week or so.

 

Oh look, a faulty gpu!