Intel bug incoming? Meltdown and Spectre exploits

Hotwired · 22 May 2018 at 14:21

Rroff said:
There are a few posters here who will deny, deny, deny, deflect, personal attack if you even hint there might be a problem AMD wise - even if AMD themselves put in black and white on their web-site :s

On the Intel side unfortunately if you are doing stuff like VMs, etc. where security is a concern then the OS patches and microcode updates just aren't optional but for a gaming PC assuming it is just normal gaming and a bit of casual browsing, etc. aside from needing browser mitigations there is limited exposure to remote intrusion via these exploits.

What we need is a thread where all the solid and not imaginary facts of how AMD is affected can be discussed just like this one that Intel has.

Rroff you present yourself as someone with the knowledge, could you start such a thread.

Rroff · 22 May 2018 at 14:21

ltron said:
However, I read an article that suggested that the currently optional microcode updates will eventually be made mandatory and automatically distributed through Windows Update. So I'll wait and see what happens.

Yay Windows 10

one of the reason I stick with 7 where possible on any machine I actually want to use.

Hotwired said:
Rroff you present yourself as someone with the knowledge, could you start such a thread.

I assume you saw how the last one went...

Ad_Augendae · 23 May 2018 at 18:57

Ad_Augendae said:

After applying the patch i did a few passes just to make sure what I was seeing wasn't skewed by something in the background.

Before

Code:

* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

   Sequential Read (Q= 32,T= 1) :  1685.553 MB/s
  Sequential Write (Q= 32,T= 1) :   594.327 MB/s
  Random Read 4KiB (Q=  8,T= 8) :   659.853 MB/s [ 161096.9 IOPS]
 Random Write 4KiB (Q=  8,T= 8) :   587.724 MB/s [ 143487.3 IOPS]
  Random Read 4KiB (Q= 32,T= 1) :   251.397 MB/s [  61376.2 IOPS]
 Random Write 4KiB (Q= 32,T= 1) :   199.730 MB/s [  48762.2 IOPS]
  Random Read 4KiB (Q=  1,T= 1) :    36.348 MB/s [   8874.0 IOPS]
 Random Write 4KiB (Q=  1,T= 1) :    88.460 MB/s [  21596.7 IOPS]

  Test : 1024 MiB [C: 19.0% (88.2/464.1 GiB)] (x5)  [Interval=5 sec]
  Date : 2018/05/16 11:06:05
    OS : Windows 10  [10.0 Build 17134] (x64)

After

Code:

* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

   Sequential Read (Q= 32,T= 1) :  1707.491 MB/s
  Sequential Write (Q= 32,T= 1) :   192.551 MB/s
  Random Read 4KiB (Q=  8,T= 8) :   648.985 MB/s [ 158443.6 IOPS]
 Random Write 4KiB (Q=  8,T= 8) :    46.129 MB/s [  11262.0 IOPS]
  Random Read 4KiB (Q= 32,T= 1) :   261.102 MB/s [  63745.6 IOPS]
 Random Write 4KiB (Q= 32,T= 1) :    26.257 MB/s [   6410.4 IOPS]
  Random Read 4KiB (Q=  1,T= 1) :    34.924 MB/s [   8526.4 IOPS]
 Random Write 4KiB (Q=  1,T= 1) :     0.974 MB/s [    237.8 IOPS]

  Test : 1024 MiB [C: 19.0% (88.2/464.1 GiB)] (x5)  [Interval=5 sec]
  Date : 2018/05/16 21:16:31
    OS : Windows 10  [10.0 Build 17134] (x64)

Hotwired said:
That's quite a hit.

I'm happy to report after messing about with Bios etc I'm back upto snuff. Another episode of huh!

Code:

* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

   Sequential Read (Q= 32,T= 1) :  1665.036 MB/s
  Sequential Write (Q= 32,T= 1) :   590.962 MB/s
  Random Read 4KiB (Q=  8,T= 8) :   664.422 MB/s [ 162212.4 IOPS]
 Random Write 4KiB (Q=  8,T= 8) :   586.637 MB/s [ 143221.9 IOPS]
  Random Read 4KiB (Q= 32,T= 1) :   248.119 MB/s [  60575.9 IOPS]
 Random Write 4KiB (Q= 32,T= 1) :   202.164 MB/s [  49356.4 IOPS]
  Random Read 4KiB (Q=  1,T= 1) :    35.851 MB/s [   8752.7 IOPS]
 Random Write 4KiB (Q=  1,T= 1) :    88.043 MB/s [  21494.9 IOPS]

  Test : 1024 MiB [C: 14.9% (69.2/464.1 GiB)] (x5)  [Interval=5 sec]
  Date : 2018/05/23 18:52:10
    OS : Windows 10  [10.0 Build 17134] (x64)

pete910 · 23 May 2018 at 19:04

@Ad_Augendae What did you change ?

Ad_Augendae · 23 May 2018 at 19:05

A bios update

pete910 · 23 May 2018 at 19:08

Ad_Augendae said:
A bios update

That's very different to "messing with bios"

What board/bios ver ?

Ad_Augendae · 23 May 2018 at 19:12

Sorry bad expression, It's an X360 F47 to 50 (I think) It's Intel/HP machine.

pete910 · 23 May 2018 at 19:15

Ad_Augendae said:
Sorry bad expression, It's an X360 F47 to 50 (I think) It's Intel/HP machine.

Ironically the Spectre laptops .

Good it's sorted though.
We'll all have to wait and see what the next lot of mitigations bring now.

Ad_Augendae · 23 May 2018 at 19:16

I use this machine as the guinea pig for all my intel machines.

pete910 · 23 May 2018 at 19:19

Ad_Augendae said:
I use this machine as the guinea pig for all my intel machines.

Bet the update had the fixed microcode to use the software patches correctly .
They look a nice lappy tbh

Ad_Augendae · 23 May 2018 at 19:19

pete910 said:
Ironically the Spectre laptops .

Good it's sorted though.
We'll all have to wait and see what the next lot of mitigations bring now.

it's never run better! even the project function with my Roku with great sound etc, who knew ?

jigger · 15 Aug 2018 at 19:00

Ad_Augendae · 16 Aug 2018 at 12:07

Ad_Augendae said:

I'm happy to report after messing about with Bios etc I'm back upto snuff. Another episode of huh!

Code:

* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

   Sequential Read (Q= 32,T= 1) :  1665.036 MB/s
  Sequential Write (Q= 32,T= 1) :   590.962 MB/s
  Random Read 4KiB (Q=  8,T= 8) :   664.422 MB/s [ 162212.4 IOPS]
 Random Write 4KiB (Q=  8,T= 8) :   586.637 MB/s [ 143221.9 IOPS]
  Random Read 4KiB (Q= 32,T= 1) :   248.119 MB/s [  60575.9 IOPS]
 Random Write 4KiB (Q= 32,T= 1) :   202.164 MB/s [  49356.4 IOPS]
  Random Read 4KiB (Q=  1,T= 1) :    35.851 MB/s [   8752.7 IOPS]
 Random Write 4KiB (Q=  1,T= 1) :    88.043 MB/s [  21494.9 IOPS]

  Test : 1024 MiB [C: 14.9% (69.2/464.1 GiB)] (x5)  [Interval=5 sec]
  Date : 2018/05/23 18:52:10
    OS : Windows 10  [10.0 Build 17134] (x64)

Ummm....

Code:

* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

   Sequential Read (Q= 32,T= 1) :  1704.192 MB/s
  Sequential Write (Q= 32,T= 1) :   591.727 MB/s
  Random Read 4KiB (Q=  8,T= 8) :   779.910 MB/s [ 190407.7 IOPS]
 Random Write 4KiB (Q=  8,T= 8) :   588.747 MB/s [ 143737.1 IOPS]
  Random Read 4KiB (Q= 32,T= 1) :   270.778 MB/s [  66107.9 IOPS]
 Random Write 4KiB (Q= 32,T= 1) :   208.049 MB/s [  50793.2 IOPS]
  Random Read 4KiB (Q=  1,T= 1) :    36.041 MB/s [   8799.1 IOPS]
 Random Write 4KiB (Q=  1,T= 1) :    89.429 MB/s [  21833.3 IOPS]

  Test : 1024 MiB [C: 14.7% (68.0/464.1 GiB)] (x5)  [Interval=5 sec]
  Date : 2018/08/16 11:50:25
    OS : Windows 10  [10.0 Build 17134] (x64)

August 14, 2018—KB4343909 (OS Build 17134.228)

It works... \o/

NinjaCool · 23 Aug 2018 at 09:41

Intel Gags Customers from Publishing Performance Impact of Microcode Updates!
https://www.techpowerup.com/247028/...shing-performance-impact-of-microcode-updates

Intel has updated their license terms regarding their microcode update distribution to explicitly forbid its users from publishing comparative "before/after" performance numbers of patched processors :eek:

Edit: It looks like Intel will be rolling back the offending clauses in the license!

Bonjour · 23 Aug 2018 at 10:20

I can tell you.. the figures from our virtuslised SQL server pre and post Meltdown/Spectre updates were particularly bad.

humbug · 23 Aug 2018 at 11:25

NinjaCool said:
Intel Gags Customers from Publishing Performance Impact of Microcode Updates!
https://www.techpowerup.com/247028/...shing-performance-impact-of-microcode-updates

Intel has updated their license terms regarding their microcode update distribution to explicitly forbid its users from publishing comparative "before/after" performance numbers of patched processors

$220bn company has been making CPU's with more holes in them than a Tea Bag for more than a decade, instead of apologising for the performance hit their customers get from fixing their broken CPU's Intel threatens them to shut up about it or we will sue you.

Such a competent open and honest company, Intel. And you wonder why so many people don't like you much.

AndreiD · 23 Aug 2018 at 11:36

Gonna guess that the EULAs got mixed up there, because they know people are going to test it anyway since they can't really legally enforce EULAs.

Bonjour · 23 Aug 2018 at 11:41

I wonder how much closer piledriver would have been to Intel's performance without these bugs.

CAT-THE-FIFTH · 23 Aug 2018 at 11:43

NinjaCool said:
Intel Gags Customers from Publishing Performance Impact of Microcode Updates!
https://www.techpowerup.com/247028/...shing-performance-impact-of-microcode-updates

Intel has updated their license terms regarding their microcode update distribution to explicitly forbid its users from publishing comparative "before/after" performance numbers of patched processors

Wait,isn't that what I posted in the last few pages - TPU picked up on it??

Edit!!

It was in this thread:

https://forums.overclockers.co.uk/threads/foreshadow.18828376/

humbug · 23 Aug 2018 at 11:43

AndreiD said:
Gonna guess that the EULAs got mixed up there, because they know people are going to test it anyway since they can't really legally enforce EULAs.

They can't, you're right, i said that earlier, however for Intel that's not the point, all they need is the threat and people will fall into line, its a message, we are serious about wanting you to shut the _____ up!