This dont look good, Intel design flaw reduce performance

[SCG_Fenris_Wolf]

And little of this has anything to do with the CPU vulnerability or the performance impacts associated with the "fixes"...

Yep, all that went off topic. Stated that earlier already. Did you guys install the updates by now?

 

As dburne, got no difference. i7 7700k here.

Edited January 15, 2018 by 4./JG52_Fenris_Wolf

[TheSNAFU]

I am holdimg off till I see some results on comparable rigs to mine. Almost no info available so far. Thanks for sharing your results. Glad it didn't impact you or dburne.

[DaveP63]

No appreciable difference here.

[TheSNAFU]

No appreciable difference here.

 

Thanks  DaveP63, what OS are you using? Can you clarify "no appreciable difference" in terms of FPS and smoothness in BOX?

Edited January 15, 2018 by TheSNAFU

[ZachariasX]

No appreciable difference here.

There shouldn‘t be. What is affected is your SSD I/O speed, in other words you should meter loading times of the maps/missions and compare those. Furthermore, what is really affected is the exchange between user threads (to put it in a very, very coarse way), this implies that your „single user usage“ (such as your gaming box) is not affected primarily, but if you ran 100 users on your system, things would look much worse.

 

Besides, you can expect further microcode updates from Intel (from AMD as well), as thus first one tended to brick older systems and send the CPUs in perpetual reboot. Just to illustrate the care that is taken at Intel for „fixing“ their CPUs.

[JonRedcorn]

At this point, these bugs are not terribly concerning for current desktop computers. The performance losses are small, and those who cannot afford them can easily skip any updates. 

 

However, the future of processors in general is quite worrying. Many people had been hoping for incremental IPC and clock improvements to help with software such Il-2. The recent bugs might be a serious setback. I imagine that unreleased products will still be affected for quite some time. 

 

With the next generation of VR devices seemingly around the corner, any stagnation in CPU performance is very sad. I hope that the industry will be able to deliver improved products in this critical time. The software that we love is always very close to the edge of the performance envelope. 

The problem is devs need to stop coding with ancient techniques, the new 8700k has 6 cores and 12 threads and bos utilizes a total of 1. It's a joke. The cpu power is here right now, devs just aren't utilizing it.

[BeastyBaiter]

Very true and it's a massive pet peeve of mine. But it's also OT.

[DaveP63]

Thanks  DaveP63, what OS are you using? Can you clarify "no appreciable difference" in terms of FPS and smoothness in BOX?

Win 10. No difference at all in smoothness. Loss of 1-2 FPS under normal conditions winter map.

[Urra]

The problem is devs need to stop coding with ancient techniques, the new 8700k has 6 cores and 12 threads and bos utilizes a total of 1. It's a joke. The cpu power is here right now, devs just aren't utilizing it.

Waiting to see what 3.0 release will bring. But don't have really high hopes. Amd will most likely release a threadripper cpu with 32 physical cores, 64 threads in q4 2018. Wonder if well be able to use more than 2 or 4 by then.

[DaveP63]

There shouldn‘t be. What is affected is your SSD I/O speed, in other words you should meter loading times of the maps/missions and compare those. Furthermore, what is really affected is the exchange between user threads (to put it in a very, very coarse way), this implies that your „single user usage“ (such as your gaming box) is not affected primarily, but if you ran 100 users on your system, things would look much worse.

 

Besides, you can expect further microcode updates from Intel (from AMD as well), as thus first one tended to brick older systems and send the CPUs in perpetual reboot. Just to illustrate the care that is taken at Intel for „fixing“ their CPUs.

Don't have an SSD.

[ZachariasX]

The problem is devs need to stop coding with ancient techniques, the new 8700k has 6 cores and 12 threads and bos utilizes a total of 1. It's a joke. The cpu power is here right now, devs just aren't utilizing it.

This means going Vulcan or DX12 and rewriting a huge part of the som engine just as it was brought to DX11? Ouch. If they went for DX11, they probably dump half their player base as Win7 is still popular. Going Vulcan has probably less customer culling, but still sets the bar higher for a decent GPU capable of running the game.

Don't have an SSD.

You‘re lucky that your drive is so slow then that overhead by the patch are much too small to bottleneck your HD‘s performance. I wouldn‘t expect any difference with or without patch on your (single user) system.

[=362nd_FS=Hiromachi]

 

 

Vulcan

I've asked about utilizing Vulcan on Q&A over a year ago, when we were getting that Dx 11 and answer was basically no with some arguments against it. Dx 12 is a waste of time and is no good API at all, just Mircrosoft filler. Now competition is implementing Vulcan and we will see how it works there. Personally I've always expected Dx 11 here to be a temporary solution, but even inside of Dx 11 you can take advantage of multi-core cpus. 

[TheSNAFU]

 

 

Win 10. No difference at all in smoothness. Loss of 1-2 FPS under normal conditions winter map.

 

 

Thanks again DaveP63. Other than Win 10 your system is not that different from mine. I would like to see results with similar rigs on Win 7 since that's where I keep reading of big performance hits. 

[FuriousMeow]

The problem is devs need to stop coding with ancient techniques, the new 8700k has 6 cores and 12 threads and bos utilizes a total of 1. It's a joke. The cpu power is here right now, devs just aren't utilizing it.

 

Where are you getting this from? BoS takes advantage of multi-core - that's why quad core is the minimum recommended, physical not threads/virtual because those aren't any benefit for games.

 

If you mean DX12 using multi-core support for GPU processing, good luck on that since very few new games are even coding to that.

[BeastyBaiter]

BoS uses 4 threads, however it only uses 1 for the graphics engine, which is the most heavily taxed. HT/SMT is useful in games but caution is needed. HT/SMT isn't helpful when all the threads are doing the exact same types of commands (e.g. nothing but floating point math) but if doing different things, then each virtual core can almost equal a physical core. It depends on the workload. As an example, if you have one physical core doing a bunch of if/else commands (e.g. AI) on one thread and a bunch of floating point math (e.g. physics) on another, it's almost as if you have 2 physical cores. But if you have 2 threads doing floating point math on the same core, there isn't a benefit over throwing it all into a single thread.

[FuriousMeow]

I've looked into HT for games where they were tested, and there isn't enough idle time - especially - in flight/aircombat sims to use HT. Other games, shooters for example, have been tested and no benefit has been found.

 

Only DX12 (and I guess Vulcan, which was slower in DOOM for nVidia cards) takes advantage of multi-core processing for graphics, as I stated, which is still in it's infancy and for the most part not being coded to yet.

Edited January 19, 2018 by FuriousMeow

[BeastyBaiter]

Doom is much faster in Vulkan than OpenGL with both Nvidia and AMD cards. Proof:

 

Vulkan:

 

 

 

OpenGL:

 

[dburne]

I have had HT disabled on my rig just about since I built it. Never saw a benefit of it , and it just increases the temps on the cores.

[BeastyBaiter]

It's primarily useful for multi-tasking. Prime example: rendering an image with the CPU while browsing the web or playing a game. The types of commands rendering uses are totally different from web browsing or playing a game. Thus I can have my CPU rendering on 6 threads while doing whatever else on the other 6 with an unimportant loss of performance (still above 60 fps in game, no noticeable impact on browsing, no more than maybe a 1% increase in render time). On the other hand, the rendering time with 12 threads == rendering time with 6 threads.

[ZachariasX]

Linus is not so happy with Intel's CPU patches.

 

Who would have thought? They have something that works for them for two decades, then once it came public that it is maybe not so good of a feature, at least not in the way Intel implemented it, they supposedly "fix it" in two weeks time. "Two weeks, be sure!"™

 

Linuses reply on the issue the TL;DR, here:

 

............

As it is, the patches are COMPLETE AND UTTER GARBAGE.

They do literally insane things. They do things that do not make
sense. That makes all your arguments questionable and suspicious. The
patches do things that are not sane.

WHAT THE F*CK IS GOING ON?

.............

 

If you really think that cloud providers will tolerate the IBRS overhead they earned with the patch long term, then you can hold on to your Intel shares.

[SAS_Storebror]

For those who suffer from serious performance and/or stability impacts after installing the Meltdown Patch by Microsoft, this might be interesting:
https://support.microsoft.com/en-us/help/4073119/protect-against-speculative-execution-side-channel-vulnerabilities-in

Currently I have rolled back the patch and blocked it from being installed further on, but since that will likely keep me from receiving critical patches in the future, I'm happy to see that it's possible to install yet disable the patch.

Cheers!
Mike

[ZachariasX]

Spectre is dead! Long live Spectre-NG!

 

http://www.guru3d.com/news-story/eight-new-spectre-variant-vulnerabilities-for-intel-discovered-four-of-them-critical.html

 

You get eight flaws for the name of one! 4 critical! After they took the SDD speed for the last ones, I wonder what they come up with to sacrify this time. L1 and L2 cache? How long will people be stupid enough to store their data on some other peoples computer („cloud“)?