A Massive Intel Hardware Bug May Be on the Horizon

MartinX

One Hour Martinizing While You Wait
Joined
Jan 23, 2003
Messages
7,187
I'm curious to see how quickly intel can respond to this, as in, release products without the hardware fault. Depending on it's exact nature and where they are in development of their current pipeline of products, it could be quite a while.

They are probably boned if they don't have something ready for the next time Google or Amazon or MS is updating their data centres (I mean, presumably that's happening on a rolling basis fairly constantly, but the supply contracts and stuff are probably less frequent)
 
Last edited:

Orddie

2[H]4U
Joined
Dec 20, 2010
Messages
2,805
TL,DR OK, it's a security bug, but what's the symptom? What if I don't want it to be fixed? What if I want my 30% performance, security be damned?
thats your single minded view. your PC will crush everything due to youe lack of security practice but everything you acess will be slow as fuck. so in the end... you gain nothing IMO.
 
Joined
Dec 7, 2010
Messages
701
So what's the alternative for somebody who wants to build a mini-itx system?
It looks like all of AMD's AM4 mini-itx boards are complete garbage.
At this point maybe I should just start investing in an ARM SOC...
 

OrangeKhrush

[H]ard|Gawd
Joined
Dec 15, 2016
Messages
1,622
So what's the alternative for somebody who wants to build a mini-itx system?
It looks like all of AMD's AM4 mini-itx boards are complete garbage.
At this point maybe I should just start investing in an ARM SOC...
of course you can always consider non ITX builds?

There is a ROG and Fatality ITX board for AM4, I am sure they are not garbage.
 

/dev/null

[H]F Junkie
Joined
Mar 31, 2001
Messages
14,711
with the Ryzen gcc compiler bug in which (last I read) didn't have an exact silicon revision with a fix, and the Intel pti bug, I think the only logical choice at this point is Threadripper.
 

M76

[H]F Junkie
Joined
Jun 12, 2012
Messages
11,161
thats your single minded view. your PC will crush everything due to youe lack of security practice but everything you acess will be slow as fuck. so in the end... you gain nothing IMO.
What?
 

/dev/null

[H]F Junkie
Joined
Mar 31, 2001
Messages
14,711
Last reports were that AMD fixed the GCC bug and even replaced the chips affected by it.
True, with newer silicon revisions, but what I don't want to to is head 1.5 hrs to microcenter, and then have to either RMA a brand new cpu when it fails or drive back to MC to get a replacement. Until there is an exact silicon stepping I can reference (eg: week 26 or later) or all broken are pulled from the shelves, I'm not playing the lottery.
 

OrangeKhrush

[H]ard|Gawd
Joined
Dec 15, 2016
Messages
1,622
True, with newer silicon revisions, but what I don't want to to is head 1.5 hrs to microcenter, and then have to either RMA a brand new cpu when it fails or drive back to MC to get a replacement. Until there is an exact silicon stepping I can reference (eg: week 26 or later) or all broken are pulled from the shelves, I'm not playing the lottery.
I am thinking that kind of bug can be more akin to new architectures and a somewhat rushed release, it was a bit of a teething issue, a nuisance for sure but fixable, security issues are different, especially if people know things they aren't supposed to.
 

/dev/null

[H]F Junkie
Joined
Mar 31, 2001
Messages
14,711
I am thinking that kind of bug can be more akin to new architectures and a somewhat rushed release, it was a bit of a teething issue, a nuisance for sure but fixable, security issues are different, especially if people know things they aren't supposed to.
By the time I troubleshoot several hours, correspond with AMD, drive to drop off usps/fedex/ups package for RMA, I'd be much better off with an 8c/16t threadripper that will work out of the box & be less frustrated as well.
At this point a good x370 board is what, $100-$150 less than a basic x399 board & the cpu difference is like $100? Rather spend my money on TR (better platform) than my time (time is money) for the same cost.

TR > Ryzen > Intel at this point with the bugs known over the last 9 months or so, IMHO.
 

ghostwich

[H]ard|Gawd
Joined
Sep 10, 2014
Messages
1,749

drescherjm

[H]F Junkie
Joined
Nov 19, 2008
Messages
14,806
True, with newer silicon revisions, but what I don't want to to is head 1.5 hrs to microcenter, and then have to either RMA a brand new cpu when it fails or drive back to MC to get a replacement. Until there is an exact silicon stepping I can reference (eg: week 26 or later) or all broken are pulled from the shelves, I'm not playing the lottery.
That has been very frustrating for me because I want to build a linux based Ryzen 1700 server / pvr. Also week 25 is not a guaranteed fix if you believe the thread at AMD. There were still users reporting the problem in December having purchased their CPU in December. At this point I will probably just wait till Q2 and pick up one of the new cpus.
 
Joined
Dec 7, 2010
Messages
701
That has been very frustrating for me because I want to build a linux based Ryzen 1700 server / pvr. Also week 25 is not a guaranteed fix if you believe the thread at AMD. There were still users reporting the problem in December having purchased their CPU in December.
Sigh.
 
  • Like
Reactions: erek
like this

dook43

2[H]4U
Joined
Sep 9, 2005
Messages
2,967
True, with newer silicon revisions, but what I don't want to to is head 1.5 hrs to microcenter, and then have to either RMA a brand new cpu when it fails or drive back to MC to get a replacement. Until there is an exact silicon stepping I can reference (eg: week 26 or later) or all broken are pulled from the shelves, I'm not playing the lottery.
My launch chip doesn't have the issue.
 
  • Like
Reactions: erek
like this

dook43

2[H]4U
Joined
Sep 9, 2005
Messages
2,967
By the time I troubleshoot several hours, correspond with AMD, drive to drop off usps/fedex/ups package for RMA, I'd be much better off with an 8c/16t threadripper that will work out of the box & be less frustrated as well.
At this point a good x370 board is what, $100-$150 less than a basic x399 board & the cpu difference is like $100? Rather spend my money on TR (better platform) than my time (time is money) for the same cost.

TR > Ryzen > Intel at this point with the bugs known over the last 9 months or so, IMHO.
Doesn't the 1900x actually have less performance in certain workloads (gaming) due to the core organization (2 1600X glued together)?
 

drescherjm

[H]F Junkie
Joined
Nov 19, 2008
Messages
14,806
By the time I troubleshoot several hours, correspond with AMD, drive to drop off usps/fedex/ups package for RMA, I'd be much better off with an 8c/16t threadripper that will work out of the box & be less frustrated as well.
At this point a good x370 board is what, $100-$150 less than a basic x399 board & the cpu difference is like $100? Rather spend my money on TR (better platform) than my time (time is money) for the same cost.
I have considered that. However I would prefer the Ryzen over the TR for one important reason idle power. The last time I checked TR had a much higher idle power and for me this is important because the sever will be a 24/7/365 box.




With that said I am highly disappointed with the Intel bug that this thread was supposed to talk about. My alternate option (if Ryzen did not fix the problem) was to get 6C/12T E3 v7 xeon (yes I know they are not out yet and not even sure there is an ETA)

This bug would make that option less desirable unless I could disable the OS fix. For my needs I am the only user of the server I don't need to protect me from myself.
 
Last edited:

longblock454

[H]ard|Gawd
Joined
Nov 28, 2004
Messages
1,790
This bug would make that option less desirable unless I could disable the OS fix. For my needs I am the only user of the server I don't need to protect me from myself.
You can on Linux with the nopti Kernel boot parameter. No idea on Windows.
 

Meeho

Supreme [H]ardness
Joined
Aug 16, 2010
Messages
4,989
I've literally built a new i7-8700K rig yesterday, after deliberating between it and Ryzen for a couple of weeks. Happy new year Intel owners!
 

GDI Lord

Limp Gawd
Joined
Jan 14, 2017
Messages
201
Possibly, but I'd prefer to think that it was just an honest mistake in an /incredibly/ complex system.
 
  • Like
Reactions: ole-m
like this

/dev/null

[H]F Junkie
Joined
Mar 31, 2001
Messages
14,711
I have considered that. However I would prefer the Ryzen over the TR for one important reason idle power. The last time I checked TR had a much higher idle power and for me this is important because the sever will be a 24/7/365 box.




With that said I am highly disappointed with the Intel bug that this thread was supposed to talk about. My alternate option (if Ryzen did not fix the problem) was to get 6C/12T E3 v7 xeon (yes I know they are not out yet and not even sure there is an ETA)

This bug would make that option less desirable unless I could disable the OS fix. For my needs I am the only user of the server I don't need to protect me from myself.
Thx for the graph! In my case, I pay about $1/w/year if left on 24/7 so that is about $45/year idle. Don't really care at that price...
 

Grimlaking

2[H]4U
Joined
May 9, 2006
Messages
3,250
This has nothing to do with virtualization other than it also utilizes virtual memory. Every program uses virtual memory.
Really that would be a shock to my ESXI hosts running guests consuming a lot less memory than allocated either hardware wise to the host or allocated to the guests running on that host. In windows yes... we always run SOME sort of virtual memory. Maybe this is where NVME and Octane will come in to play?
 

Grimlaking

2[H]4U
Joined
May 9, 2006
Messages
3,250
this is a very complex issue Ive been reading into it for most of the night it is a lot ot digest TLDR wait for the embargo to be up then is when the shit will meet the fan just how much shit i can speculate (Mexico City's yearly output) it has to be a huge issue for all cloud providers to be rebooting all their systems between this friday and next wednesday.
And that is only if the hosting software and guests are all patched in that window. Just because the embargo lifts doesn't mean the patch will be ready across the board. I agree though this will be an interesting one to follow.
 

Verado

Limp Gawd
Joined
May 16, 2017
Messages
266
I wish it didn't exist. Up to 35% = up to three threads of my i7-6700.
Yeah this is a bummer.
With the 8 gen also affected i currently have no viable upgrade paths as i am all about single core performance memory bandwidth. So ryzen is out of the picture for me.
 

jbc029

Limp Gawd
Joined
Mar 26, 2013
Messages
484
F*ck me. A 10% hit in database performance will hurt...a lot. That actually hits me where I work, and I don't like that. :(
 

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
5,194
So how will this affect Netflix, Amazon.com, Banks, Facebook, Twitter, Steam, Xbox Live, . . . -> Basically what people use on the internet? While local processes like gaming unless MP is involved maybe not much at all. Will we see service crashes on some services due to this rolling out and the decrease capacity of the servers? Man this came out of the blue, no pun intended.
 

Mohonri

Supreme [H]ardness
Joined
Jul 29, 2005
Messages
5,751
So here's my question: how long will it take for Intel to fix the hardware bug, prototype/test it, and get new CPUs to market? I imagine that between now and then, their corporate/business sales are going to tank badly--businesses are going to want to defer purchases until after new CPUs are available. Or they'll opt for Ryzen/Threadripper.
 

Verado

Limp Gawd
Joined
May 16, 2017
Messages
266
So here's my question: how long will it take for Intel to fix the hardware bug, prototype/test it, and get new CPUs to market? I imagine that between now and then, their corporate/business sales are going to tank badly--businesses are going to want to defer purchases until after new CPUs are available. Or they'll opt for Ryzen/Threadripper.
And will the hardware fix come at a performance cost? Not likely though.
 
Top