[Various] Spectre & Meltdown: Critical vulnerabilities in modern processors - Page 3 - Overclock.net - An Overclocking Community

Forum Jump: 

[Various] Spectre & Meltdown: Critical vulnerabilities in modern processors

Reply
 
Thread Tools
post #21 of 1852 (permalink) Old 01-02-2018, 02:04 PM
Waiting for 7nm EUV
 
tpi2007's Avatar
 
Join Date: Nov 2010
Posts: 11,307
Rep: 893 (Unique: 502)
Quote:
Originally Posted by Avonosac View Post

Quote:
Originally Posted by tpi2007 View Post

Warning: Spoiler! (Click to show)
I agree with them posting Intel numbers.

When it comes to AMD, that part has perhaps the unintended consequence that people will know that something happened behind the scenes if AMD CPUs also degrade in performance when they shouldn't.

The correct way to post that story would be:

1. Here are the Intel numbers;
2. Here are the AMD numbers if something behind the scenes happens.

I disagree completely with any association with AMD. This is disclosed and referenced as a solely *INTEL* bug. If at any point in the future something in this affects AMD directly somehow, then and *ONLY* then do they even get mentioned. The only exception being the statement AMD is unaffected by this bug and has no performance impact.

This legitimately may make EPYC superior in performance to intel, not just better PPD. The new normal is to compare AMD throughput on all tasks to Intel's new throughput when the fixes are released to all kernels.

I was referring to AMD's reply to someone's opinion that:
Quote:
/* Assume for now that ALL x86 CPUs are insecure */
- setup_force_cpu_bug(X86_BUG_CPU_INSECURE);
+ if (c->x86_vendor != X86_VENDOR_AMD)
+ setup_force_cpu_bug(X86_BUG_CPU_INSECURE);

fpu__init_system(c);

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


Edit: Just to be clear, I agree that without trying to source Intel numbers for the article, they do seem incompetent. It's hard to say that it's outright bad faith because you'd somehow have to selectively not read parts of the article both before and after that part with the AMD numbers tweet to not understand that AMD CPUs are not affected.

Is your CPU bottlenecking your GPU ? Find out: CPU and GPU usage along with FPS in-game
Read my reviews here.
Clubs (founder): The rare / unusual CPU club
Clubs (member): Corsair Professional HX / AX Series PSU Owners Club || The Official Cooler Master HAF X/932/922/912(+) Club
CPU
Core i7-3820
Motherboard
Asus Sabertooth X79
GPU
MSI GTX 1060 6 GB Gaming X
RAM
16 GB Corsair DDR3 1866 Mhz Dominator
Hard Drive
Samsung SSD 830 128GB + WD Caviar Black 2TB
Optical Drive
Sony Optiarc DVD-RW
Power Supply
Corsair AX750 Professional Modular 80 Plus Gold
Cooling
Corsair A70 + Noiseblocker M12-P
Case
Cooler Master HAF 912 Plus
Operating System
Windows 7 Home Premium 64-bit
Monitor
BenQ RL2455HM
Keyboard
Cooler Master Octane
Mouse
Cooler Master Octane
▲ hide details ▲


tpi2007 is offline  
Sponsored Links
Advertisement
 
post #22 of 1852 (permalink) Old 01-02-2018, 02:37 PM
New to Overclock.net
 
Ghoxt's Avatar
 
Join Date: Dec 2009
Location: SW Virginia
Posts: 2,877
Rep: 159 (Unique: 128)

So correct me if I'm wrong...Did Intel lose control of one of the vulnerabilities they designed on purpose for 3 letter "Agencies", /tinfoil :rolleyes: :o


"Everyone is entitled to their own Opinions, but not their own Facts!" cigar.gif

Ghoxt is offline  
post #23 of 1852 (permalink) Old 01-02-2018, 02:44 PM
Newb to Overclock.net
 
mouacyk's Avatar
 
Join Date: Jan 2013
Posts: 3,867
Rep: 167 (Unique: 124)
[email protected] breaks it down (matches up with explanations on ycombinator as well):
Quote:
It sounds like this is tied to speculative execution. If you're speculatively executing an instruction then it is possible you'll just end up throwing away the result anyway, so you want to do it as cheaply as possible. Maybe Intel figured out that they can skip the priv checks while speculatively executing, and then perform them before actually implementing the results if it turns out the instruction was needed. However, maybe it turns out that the speculative execution opens up some back-door way of getting at the data, such as via the cache/timing/etc, which wouldn't be exposed if an exception was raised sooner.
Sucks. Intel's 30% performance lead is rearing it's ugly head, and the debt is due.

CPU: Q6600 X5470 2100T 1680V2 9900K
GPU: 8800GT 9800GTX+ 750Ti 1080Ti
RAM: 2x8GBTZ
Gentoo64 in Water
(14 items)
LGA775 X5470
(6 items)
CPU
9900K 5GHz 1.224v
Motherboard
EVGA Z370 Micro
GPU
MSI 1080TI GXEK 2100.5/12627
RAM
16GB Trident Z 4000C16
Hard Drive
970 EVO 500GB
Power Supply
Seasonic X850 Gold
Cooling
480mm Radiator Custom
Case
Silverstone FT03
Operating System
Windows 7 Ultimate 64-bit
Operating System
Gentoo Linux 64 Multi-Lib
Monitor
Acer Predator XB271UH 165Hz
Keyboard
Logitech G710+
Mouse
Logitech G502
Audio
Sound Blaster Z
CPU
X5470 4GHz (stock v)
Motherboard
GA-EP45-UD3P
GPU
EVGA 9800 GTX+ 512MB
RAM
8GB 4x2GB GSkill 1066MHz DDR2
Cooling
XSPC Rasa, D5 + Res, 240mm Rad
Case
Lian-Li PC7-HX
▲ hide details ▲
mouacyk is offline  
Sponsored Links
Advertisement
 
post #24 of 1852 (permalink) Old 01-02-2018, 02:55 PM
Waiting for 7nm EUV
 
tpi2007's Avatar
 
Join Date: Nov 2010
Posts: 11,307
Rep: 893 (Unique: 502)
Quote:
Originally Posted by Ghoxt View Post

So correct me if I'm wrong...Did Intel lose control of one of the vulnerabilities they designed on purpose for 3 letter "Agencies", /tinfoil rolleyes.gif  redface.gif

You're probably wrong because those are usually software bugs where plausible deniability means that you'll never know if it was incompetence or on purpose, and the fix does not entail significant performance loss, not on a hardware problem that, if confirmed, will mean that their CPUs will lose a not insignificant amount of performance when the OSes are patched for security. That's many millions, potentially billions of dollars worth of a mistake, so the simplest explanation is that this was a mistake, an oversight in the design.


Quote:
Originally Posted by mouacyk View Post

[email protected] breaks it down (matches up with explanations on ycombinator as well):
Quote:
It sounds like this is tied to speculative execution. If you're speculatively executing an instruction then it is possible you'll just end up throwing away the result anyway, so you want to do it as cheaply as possible. Maybe Intel figured out that they can skip the priv checks while speculatively executing, and then perform them before actually implementing the results if it turns out the instruction was needed. However, maybe it turns out that the speculative execution opens up some back-door way of getting at the data, such as via the cache/timing/etc, which wouldn't be exposed if an exception was raised sooner.
Sucks. Intel's 30% performance lead is rearing it's ugly head, and the debt is due.

It's kind of ironically fitting that in a world that is speeding up with cheap speculations of all sorts that this would happen. Let's hope that those AIs that make their own code don't choose to speculate cheaply too much. redface.gif

Is your CPU bottlenecking your GPU ? Find out: CPU and GPU usage along with FPS in-game
Read my reviews here.
Clubs (founder): The rare / unusual CPU club
Clubs (member): Corsair Professional HX / AX Series PSU Owners Club || The Official Cooler Master HAF X/932/922/912(+) Club
CPU
Core i7-3820
Motherboard
Asus Sabertooth X79
GPU
MSI GTX 1060 6 GB Gaming X
RAM
16 GB Corsair DDR3 1866 Mhz Dominator
Hard Drive
Samsung SSD 830 128GB + WD Caviar Black 2TB
Optical Drive
Sony Optiarc DVD-RW
Power Supply
Corsair AX750 Professional Modular 80 Plus Gold
Cooling
Corsair A70 + Noiseblocker M12-P
Case
Cooler Master HAF 912 Plus
Operating System
Windows 7 Home Premium 64-bit
Monitor
BenQ RL2455HM
Keyboard
Cooler Master Octane
Mouse
Cooler Master Octane
▲ hide details ▲


tpi2007 is offline  
post #25 of 1852 (permalink) Old 01-02-2018, 04:04 PM
New to Overclock.net
 
sugarhell's Avatar
 
Join Date: Apr 2012
Location: 38.051289,23.709791
Posts: 9,327
Rep: 564 (Unique: 287)
This doesnt look good.

https://www.phoronix.com/scan.php?page=article&item=linux-415-x86pti&num=2

Quote:
Originally Posted by BiG StroOnZ
Warning: Spoiler! (Click to show)
No I said if Vega is clocked @ 1600MHz out of the box, I will eat my shoe on Twitch.tv.

Sound good? wink.gif
Warning: Spoiler! (Click to show)
Quote:
Originally Posted by Juicin View Post
Not in games, which is what 99.9999 percent of us care about

And unless you own a business, you get paid for your time. So your compute power just makes more work for you not less

Ryzen is a joke to the vast majority of the market who would consider buying their product.
sugarhell is offline  
post #26 of 1852 (permalink) Old 01-02-2018, 04:36 PM
Fantastic Mr Fox
 
randomizer's Avatar
 
Join Date: Apr 2009
Posts: 5,814
Rep: 175 (Unique: 120)
Quote:
Originally Posted by tpi2007 View Post

I was referring to AMD's reply to someone's opinion that:
https://lkml.org/lkml/2017/12/27/2

This patch never happened. AMD CPUs are impacted in the current kernel source. Here's the commit:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/x86/kernel/cpu/common.c?id=a89f040fa34ec9cd682aed98b8f04e3c47d998bd

CPU
i7 920 D0
Motherboard
MSI X58 Pro-E
GPU
Gigabyte GTX 970 (GV-N970IX-4GD)
RAM
3x2GB G.Skill DDR3-1333 9-9-9-24
Hard Drive
840 Pro
Hard Drive
Caviar Black
Optical Drive
LG BD-ROM
Power Supply
Corsair HX-520
Case
CM690
Operating System
Windows 8.1 Pro x64
Monitor
Dell U2713HM
Monitor
Dell U2311H
Keyboard
Turbo-Trak (Google it :D)
Mouse
Mionix Avior 7000
Mouse
Everglide Titan
Audio
AKG K 242 HD
▲ hide details ▲
randomizer is offline  
post #27 of 1852 (permalink) Old 01-02-2018, 05:21 PM
Waiting for 7nm EUV
 
tpi2007's Avatar
 
Join Date: Nov 2010
Posts: 11,307
Rep: 893 (Unique: 502)
Quote:
Originally Posted by sugarhell View Post

This doesnt look good.

https://www.phoronix.com/scan.php?page=article&item=linux-415-x86pti&num=2

Syntehtics look quite bad for Coffe Lake compared to Broadwell-E (between -38% and -54% after the patch). Probably applies to all Skylake based CPUs (Skylake, Kaby Lake and Coffee Lake). Let's see if it translates that badly in some real life applications. The 8700K is still faster than the 6800K at the end of the day, but how does it compare to Ryzen without the same patch? That'll be interesting to see.

Compile Bench is ~14.8% down for both Coffee Lake and Broadwell-E.
H.264 video encoding is within margin of error (less than 1% for both systems), same for Linux Kernel compilation and FFmpeg transcoding.

It seems, at least for now, as they say, "applications mostly limited to user-space activity should see minimal change (if any) in performance."

The problem seems to be mostly with heavy duty stuff. PostgreSQL is down ~13% for the 8700K and ~19% for Broadwell-E (still faster than the 8700K in either case though) and Redis is down ~5,8% for the 8700K and ~7,2% for the 6800K, with the 8700K still winning in either case.

He says more benchmarks are incoming. Will be interesting to follow.


Quote:
Originally Posted by randomizer View Post

Quote:
Originally Posted by tpi2007 View Post

I was referring to AMD's reply to someone's opinion that:
https://lkml.org/lkml/2017/12/27/2

This patch never happened. AMD CPUs are impacted in the current kernel source. Here's the commit:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/x86/kernel/cpu/common.c?id=a89f040fa34ec9cd682aed98b8f04e3c47d998bd

Yep, hence why it's important to mention this.

Is your CPU bottlenecking your GPU ? Find out: CPU and GPU usage along with FPS in-game
Read my reviews here.
Clubs (founder): The rare / unusual CPU club
Clubs (member): Corsair Professional HX / AX Series PSU Owners Club || The Official Cooler Master HAF X/932/922/912(+) Club
CPU
Core i7-3820
Motherboard
Asus Sabertooth X79
GPU
MSI GTX 1060 6 GB Gaming X
RAM
16 GB Corsair DDR3 1866 Mhz Dominator
Hard Drive
Samsung SSD 830 128GB + WD Caviar Black 2TB
Optical Drive
Sony Optiarc DVD-RW
Power Supply
Corsair AX750 Professional Modular 80 Plus Gold
Cooling
Corsair A70 + Noiseblocker M12-P
Case
Cooler Master HAF 912 Plus
Operating System
Windows 7 Home Premium 64-bit
Monitor
BenQ RL2455HM
Keyboard
Cooler Master Octane
Mouse
Cooler Master Octane
▲ hide details ▲


tpi2007 is offline  
post #28 of 1852 (permalink) Old 01-02-2018, 07:24 PM
Waiting for 7nm EUV
 
tpi2007's Avatar
 
Join Date: Nov 2010
Posts: 11,307
Rep: 893 (Unique: 502)
Phoronix has good news for games (at least on Linux, but should probably be the same on Windows), no impact on the games tested - CS:GO, the latest Deus Ex, Dota 2, Dawn of War III, F1 2017 and The Talos Principle):

https://www.phoronix.com/scan.php?page=news_item&px=x86-PTI-Initial-Gaming-Tests


Edit: relevant to what we were discussing above, randomizer:

https://www.phoronix.com/scan.php?page=news_item&px=x86-PTI-EPYC-Linux-4.15-Test
Quote:
For Now At Least AMD CPUs Are Also Reported As "Insecure"
Quote:
But over one week later, that patch has yet to be merged to the mainline kernel. When booting the Linux 4.15 kernel on an AMD EPYC box, indeed, for now the AMD CPU is still treated with a bug of "insecure_cpu."

An immediate workaround at least until the AMD patch lands where PTI isn't applied to AMD CPUs is by booting the kernel with the nopti kernel command-line parameter. This can also be applied to Intel systems too on a patched kernel if wanting to regain the performance and are not too concerned about this vulnerability.

Is your CPU bottlenecking your GPU ? Find out: CPU and GPU usage along with FPS in-game
Read my reviews here.
Clubs (founder): The rare / unusual CPU club
Clubs (member): Corsair Professional HX / AX Series PSU Owners Club || The Official Cooler Master HAF X/932/922/912(+) Club
CPU
Core i7-3820
Motherboard
Asus Sabertooth X79
GPU
MSI GTX 1060 6 GB Gaming X
RAM
16 GB Corsair DDR3 1866 Mhz Dominator
Hard Drive
Samsung SSD 830 128GB + WD Caviar Black 2TB
Optical Drive
Sony Optiarc DVD-RW
Power Supply
Corsair AX750 Professional Modular 80 Plus Gold
Cooling
Corsair A70 + Noiseblocker M12-P
Case
Cooler Master HAF 912 Plus
Operating System
Windows 7 Home Premium 64-bit
Monitor
BenQ RL2455HM
Keyboard
Cooler Master Octane
Mouse
Cooler Master Octane
▲ hide details ▲


tpi2007 is offline  
post #30 of 1852 (permalink) Old 01-02-2018, 08:24 PM
Iconoclast
 
Blameless's Avatar
 
Join Date: Feb 2008
Posts: 30,086
Rep: 3134 (Unique: 1869)
Quote:
Originally Posted by tpi2007 View Post

Phoronix has good news for games (at least on Linux, but should probably be the same on Windows), no impact on the games tested - CS:GO, the latest Deus Ex, Dota 2, Dawn of War III, F1 2017 and The Talos Principle)

Games don't access the kernel that much. I/O heavy tasks do.

Databases and servers are probably in trouble. Consumer tasks, not so much.

Windows users should bench their most common tasks while they have the "show kernel times" option checked in taskmgr's graphs. The more kernel time, the bigger the performance hit from a fix is likely to be.

...rightful liberty is unobstructed action according to our will within limits drawn around us by the equal rights of others. I do not add 'within the limits of the law,' because law is often but the tyrant's will, and always so when it violates the right of an individual. -- Thomas Jefferson
Blameless is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off