• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."
  • Community Question: What makes a good motherboard?

Massive security hole in CPU's incoming?Official Meltdown/Spectre Discussion Thread

Page 34 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

wahdangun

Golden Member
Feb 3, 2011
1,004
139
106
Nothing to do with OoO or IO.

Yes, but since we use vdi deployment, we just need to rebuild golden image and the server will take care the rest by issuing VM instances, when the client request it.


On another note, M$ truly want my life miserable, I mean why the hell it's good idea to update in Sunday ? And now I'm scrambling to the office fixing the server because somehow after the update my server get stuck in boot loop.
 
  • Like
Reactions: lightmanek and IEC

Zstream

Diamond Member
Oct 24, 2005
3,397
277
136
Yes, but since we use vdi deployment, we just need to rebuild golden image and the server will take care the rest by issuing VM instances, when the client request it.


On another note, M$ truly want my life miserable, I mean why the hell it's good idea to update in Sunday ? And now I'm scrambling to the office fixing the server because somehow after the update my server get stuck in boot loop.
I just mean that the thin clients could be impacted.
 

DrMrLordX

Lifer
Apr 27, 2000
16,645
5,654
136
Do you have any links that talk about Broadwell-EP and Retpoline specifically? Because I have that cpu.
I've heard from multiple sources that Retpoline won't work at all (or won't work all that well) on anything Skylake and later. So Broadwell would be the latest generation of Intel CPUs that could still take full advantage of Retpoline:

https://news.ycombinator.com/item?id=16085672

Only kinda effective. Retpoline on Skylake and later isn't completely reliable as a mitigation for one of the spectre variants, and IBRS is one of the processor features that will supposedly correct that.

You'll still be vulnerable, it'll just add another layer of difficulty to the exploit.
 

dark zero

Platinum Member
Jun 2, 2015
2,542
100
91
Intel, you want to make me migrate to AMD so fast? Despite I am expecting to see some performance penalty with the patch, I didn't expect to see more penalty than expecting...
 
  • Like
Reactions: Kuosimodo

aigomorla

Cases and Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
18,291
986
126
Safe ? Just get a Ryzen or Threadripper.......
No because a skylake-x / coffelake will still outperform both of them in gaming even with patches to bios, and also probably anything else which is more single threaded and not fully multi threaded like a lot of older programs.
 
  • Like
Reactions: hnizdo

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
20,886
9,082
136
No because a skylake-x / coffelake will still outperform both of them in gaming even with patches to bios, and also probably anything else which is more single threaded and not fully multi threaded like a lot of older programs.
And the entire world is all about gaming ? And before they were only down a few percent, and now loosing some of that, you are so confident they will still win ? A little too fanboyish if you ask me, even before the benchmarks are out.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
20,886
9,082
136
I don't think I've seen any Win10 gaming benchmarks run with OS and microcode patches.
Exactly. So who knows who wins. But Ryzen was close before, and we know that the Intel one will loose some lead, just not how much until we see benchmarks. It is actually POSSIBLE they could now be in the lead. Note I cappped the possible, not holding my breath.
 

mattiasnyc

Senior member
Mar 30, 2017
356
337
106
No because a skylake-x / coffelake will still outperform both of them in gaming even with patches to bios, and also probably anything else which is more single threaded and not fully multi threaded like a lot of older programs.
Plenty of us use multi-threaded software.

Secondly, for anyone for whom spending the maximum amount possible for a CPU is a problem there is by definition a limit, and thus the question isn't whether or not the best whatever-lake outperforms the best Ryzen, it is what the performance is per dollar at whatever amount a person is willing to spend.... and again; it's performance-per-dollar on whatever software one uses.

Black Friday in the US was of course an anomaly, but dang; Ryzen 1700 for what... about $220? Cooler included. $30 off a bundled motherboard... I'm just not seeing Intel compete with that at that price level, at the very least not on multithreaded software....
 

aigomorla

Cases and Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
18,291
986
126
I don't think I've seen any Win10 gaming benchmarks run with OS and microcode patches.
Core i3-8300:
https://www.techspot.com/article/1556-meltdown-and-spectre-cpu-performance-windows/page2.html

Battlefield 1:


Ashes of Singularity:


Assasins Creed Origins:


Rainbow Six Siege:



You see a trend here? Its almost no performance hit....
So yes im fairly confident to say win10 gaming a intel will win....

Plenty of us use multi-threaded software.

Secondly, for anyone for whom spending the maximum amount possible for a CPU is a problem there is by definition a limit, and thus the question isn't whether or not the best whatever-lake outperforms the best Ryzen, it is what the performance is per dollar at whatever amount a person is willing to spend.... and again; it's performance-per-dollar on whatever software one uses.
Whatever floats your boat is cool with me... As you said performance per dollar is what it comes out to...
OK, then lets not forget how the i5-8400 got the award of being the best cpu in gaming.
http://www.pcgamer.com/best-cpu-of-the-year-intel-core-i5-8400/

And yes we are all not gamers, however a lot of us are.
You can look here:
https://forums.anandtech.com/forums/general-hardware.4/

In a majority of the help me build a new system threads, it states gamer.
So a lot of people looking though this forum section are gamers and overclockers as well.

So that dollar to performance average can go either way for either CPU and is not only dominated by Ryzen even AFTER patch.


Honestly i wont care if your a AMD / INTEL / ARM user. I try hard to keep a neutrality possition, infact i like playing for the underdog in many occasions to help keep things neutral.

However spreading false information is not something i will tolerate.

Everyone is going off about how this patch will kill performance on said machine.
Yes that is true, but will it outrank the current standards?

Possibly on one system it may on another it might not.

We dont fully know what the full extent of what these "performance hits" are yet people are trying to make it seem like Intel has hit dirt bottom and has become a useless piece of Silicon.

If you honestly feel that way, then fine, im not judging you, however blindy telling the masses this is better then that and this is better then that without knowing the exact purpose of said machine is flat out wrong.

If your that worried about this hijack, then ok, get an AMD if that will make you feel better, but dont even try to pass info that the said AMD system will be faster then a 8700K in gaming or single threaded apps.

Also most people get a 8700K to overclock... if not why bother with the K version to begin with, so when that 8700k is hitting 5ghz... and yes most of them are hitting 5ghz + on all 6 cores, that single threaded application that some of us do happen use, the gap is still large even as well on multi threaded apps and even after all the patchs.

Those of you guys who feel like you need to change your system to a Ryzen, im not criticizing you at all, seriously, im cool with that, however those of you guys forcing others to change to ryzen because of scare tatics, that i am not down with.
 
  • Like
Reactions: Zucker2k

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
20,886
9,082
136
Core i3-8300:
https://www.techspot.com/article/1556-meltdown-and-spectre-cpu-performance-windows/page2.html

Battlefield 1:


Ashes of Singularity:


Assasins Creed Origins:


Rainbow Six Siege:



You see a trend here? Its almost no performance hit....
So yes im fairly confident to say win10 gaming a intel will win....



Whatever floats your boat is cool with me... As you said performance per dollar is what it comes out to...
OK, then lets not forget how the i5-8400 got the award of being the best cpu in gaming.
http://www.pcgamer.com/best-cpu-of-the-year-intel-core-i5-8400/

And yes we are all not gamers, however a lot of us are.
You can look here:
https://forums.anandtech.com/forums/general-hardware.4/

In a majority of the help me build a new system threads, it states gamer.
So a lot of people looking though this forum section are gamers and overclockers as well.

So that dollar to performance average can go either way for either CPU and is not only dominated by Ryzen even AFTER patch.


Honestly i wont care if your a AMD / INTEL / ARM user. I try hard to keep a neutrality possition, infact i like playing for the underdog in many occasions to help keep things neutral.

However spreading false information is not something i will tolerate.

Everyone is going off about how this patch will kill performance on said machine.
Yes that is true, but will it outrank the current standards?

Possibly on one system it may on another it might not.

We dont fully know what the full extent of what these "performance hits" are yet people are trying to make it seem like Intel has hit dirt bottom and has become a useless piece of Silicon.

If you honestly feel that way, then fine, im not judging you, however blindy telling the masses this is better then that and this is better then that without knowing the exact purpose of said machine is flat out wrong.

If your that worried about this hijack, then ok, get an AMD if that will make you feel better, but dont even try to pass info that the said AMD system will be faster then a 8700K in gaming or single threaded apps.

Also most people get a 8700K to overclock... if not why bother with the K version to begin with, so when that 8700k is hitting 5ghz... and yes most of them are hitting 5ghz + on all 6 cores, that single threaded application that some of do happen use, the gap is still large even multi threaded apps and even after all the patchs.

Those of you guys who feel like you need to change your system to a Ryzen, im not criticizing you at all, seriously, im cool with that, however those of you guys forcing others to change to ryzen because of scare tatics, that i am not down with.
They don't have Ryzen in there, but it does seem to affect performance a lot.
 

french toast

Senior member
Feb 22, 2017
978
803
106
Exactly... We need to wait until the dust clears to see the entire picture.
Yes, also that 3% might make all the difference when pinnacle ridge turns up, who knows what improvements (if any) that will bring outside of frequency.
Could run skylake very close.
 

beginner99

Diamond Member
Jun 2, 2009
4,667
1,078
136
You see a trend here? Its almost no performance hit....
So yes im fairly confident to say win10 gaming a intel will win....
Ok. Fair enough. Now we need to wait for online multi-player benches. In those the game need to send lots of small packages (IO) to the game server needing a syscall for each package. We already know that game servers receiving the packages get 3x fold CPU usage increase. So would not surprise me if it ends up to be 10-15% percent for multiplayer online-gaming.
 
  • Like
Reactions: Kuosimodo

aigomorla

Cases and Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
18,291
986
126
Ok. Fair enough. Now we need to wait for online multi-player benches. In those the game need to send lots of small packages (IO) to the game server needing a syscall for each package. We already know that game servers receiving the packages get 3x fold CPU usage increase. So would not surprise me if it ends up to be 10-15% percent for multiplayer online-gaming.
are there even benchmarks for online games like WOW / GW2 ?

Im curious to see those if any FPS benchmarks, because i know once the instance is loaded with a massive amount of players, like most MMO's it even makes the statement, "can it run crysis" under rated, if you know what i mean....

All the hits i see on google only show quality, and not a actual raw FPS count in a massive player situation.

Like for example even with a monster workhouse gaming rig, i believe my friend told me in a world event in guildwars, on 2 x titans, he could bearly maintain 30fps even on full settings, and at that point i think the system ends up being cpu limited more so then gpu limited at tracking all the other players and spells being spam'd at the world target.
 

goldstone77

Senior member
Dec 12, 2017
217
93
61
Off topic, does anyone know if any BBC code works to resize images works in these forums to make pictures smaller?
 

beginner99

Diamond Member
Jun 2, 2009
4,667
1,078
136
are there even benchmarks for online games like WOW / GW2 ?

Im curious to see those if any FPS benchmarks, because i know once the instance is loaded with a massive amount of players, like most MMO's it even makes the statement, "can it run crysis" under rated, if you know what i mean....

All the hits i see on google only show quality, and not a actual raw FPS count in a massive player situation.

Like for example even with a monster workhouse gaming rig, i believe my friend told me in a world event in guildwars, on 2 x titans, he could bearly maintain 30fps even on full settings, and at that point i think the system ends up being cpu limited more so then gpu limited at tracking all the other players and spells being spam'd at the world target.
At least for battlefield series such benchmarks exist but of course very hard to reproduce and some are still -Redacted- (empty servers). So yeah it's hard to find these benches but these would actually show the real situation because a lot of people play online games.
It also goes to show how lazy reviewers have become. They all do the same thing over and over again. Hardly anything that differentiates them or anyone going the extra mile.

Profanity is not allowed in the tech forums.

Daveybrat
AT Moderator
 
Last edited by a moderator:

DrMrLordX

Lifer
Apr 27, 2000
16,645
5,654
136
Hmm. I'd say results inconclusive on the patch + microcode update right now. The only notable losses seem to come off minfps.

It's disturbing that it's both SATA and NVMe SSDs taking a speed hit here. Sucks.
 

hnizdo

Member
Aug 11, 2017
33
16
41
I believe nvme perf hit comes after OS functioning, so it can be mitigated by MS. There is no way branch prediction can influence pcie transfers. Its MS kernel patch to blame I/O perf hit. And if its sw, it can be tuned.
 

ashetos

Senior member
Jul 23, 2013
254
14
76
I believe nvme perf hit comes after OS functioning, so it can be mitigated by MS. There is no way branch prediction can influence pcie transfers. Its MS kernel patch to blame I/O perf hit. And if its sw, it can be tuned.
Let's say that 4KB random reads QD=1 drop from 50MB/s to 40MB/s.

This means that 1 CPU core went from 78 μsec per I/O request to 97 μsec per I/O request. That is a difference of 19 μsec.

If the total overhead of meltdown patches and BIOS updates for the Windows 10 I/O path is indeed 19 μsec then it has nothing to do with PCI transfers and everything to do with CPU utilization.
 

hnizdo

Member
Aug 11, 2017
33
16
41
If the total overhead of meltdown patches and BIOS updates for the Windows 10 I/O path is indeed 19 μsec then it has nothing to do with PCI transfers and everything to do with CPU utilization.
How do you know, that 19 μsec overhead is caused by CPU, and not by OS?
 

ASK THE COMMUNITY