Return to Level1Techs.com

[LIST] of CPU's most likely immune to Spectre

cpu
hardware

#33

I see, the T3200 is based off the P6, that makes sense. Thanks!


#34

Nope the pentium T3200 is Core architecture already. Codename Merom.

But you get a heart anyway :smile_cat:


#35

I haven’t heard from authority if the Sun Sparc processors are immune to spectre, neither Oracle or Fujitsu have spoken up.

Including Sparc in a conversation like this is fun, but academic. I like Sun hardware because it is all so over-engineered. There are definite drawbacks to actually using Sun hardware.


#36

Well, at least my Atom 330 based pfSense box is safe. That’s about all the good news for me, though. Most of my current in-use stuff is almost certainly affected.

I have these systems available to test if they’d be useful:

  • Main laptop (i5-4310u)
  • Old laptop 1 (i5-2540)
  • Old laptop 2 (i5-560M)
  • Old laptop 3 (A8-3110MX)
  • Macbook Pro Early 2011 (2nd gen i5 something)
  • Main desktop (i7-6700k)
  • HTPC (i3-6100)
  • Old Desktop 1 (Phenom II X4 955 BE)
  • Old desktop 2 (Athlon II X3 435)
  • File server (Xeon E3-1220L v3)
  • VM/test server (Dual Xeon E5-2670 v1)

#37

From what I’ve read around so far, pretty much most of the mainstream Intel CPU’s post Pentium 1 are vulnerable, so any i3’s, i5’s, and i7’s (Which sucks cuz I just built a PC with a 6770k i7). Idk about the Xeon’s, Athlon’s, or Phenom’s though.
And the A8 should just be as well since it’s ARM.


#38

Well, I didn’t read the thread thoroughly; @catsay already disqualified POWER6 which does in-order execution and is still vulnerable:

This probably makes Itanium also likely to be vulnerable right? Unless someone already confirmed that and I missed that post too. Gah! Too many threads, too much information, too little time!


#39

If I recall from the info he posted on it, the Itanium runs in-order predictive branching, and Spectre/Meltdown take advantage of out-of-order predictive branching.


#40

Some fun info:

AMD’s PSP Platform Security Processor (at least from 2015 info) is an ARM Cortex A5.

Which as we know is not subject to SpecMelt.

https://www.anandtech.com/show/9319/amd-launches-carrizo-the-laptop-leap-of-efficiency-and-architecture-updates/8

Can’t say its still the case in Ryzen. But I expect not so. The requirements on Ryzen likely mean that it is using a more powerful ARM core.

Also I’ve now coined the term SpecMelt when referring to Spectre + Meltdown because the patches Melt your specs.


#41

Current RISC-V chips are immune

https://riscv.org/2018/01/more-secure-world-risc-v-isa/


#42

That puts Itanium at the same risk as POWER6 right? Both are in-order.


#43

If ultrasparc III is immune then IV should also be immune as IV is just a multi-core variant of III

T1 - T3 are also in order execution

Fujitsu implemented speculative execution on sparc64 V+ and Oracle on the T4+ so anything newer is likely vulnerable. Oracle did come out saying that many newer sparc systems are afffected, but not which ones yet.

I have some Fujitsu sparc64 VI boxes i can test on, but not T4’s


#44

Hi catsay,

First, thanks you so much for the list of CPU’s that might be immune to Spectre.

I specially registered because I wanted to thank you and let you know what I was trying to find out.

Do you remember the old Netbooks, the Asus EeePC 1000H That came with XP Home I think and had the Atom CPU N270

https://www.asus.com/nl/Laptops/Eee_PC_1000H/specifications/

Well, I was searching the whole day and even asked Asus if this old Netbook is immune to the bugs Spectre 1 and 2 and 3 and maybe even Meltdown.

They could not really give me a good answer to this only to follow the news.

I can not expect Asus to give me a BIOS update as this is a very old Netbook (Would be great but I do not think the will)

So, I got my Netbook and format it and installed Windows 10 X86 version on it ( It made it incredible slow lol, but yes it does install )

My ownly goal was to find out if this Atom has side band addressing or what makes it possible for Spectra and or the Meltdown bug ( keep confusing these 2 sorry )

Then first I download
https://www.grc.com/inspectre.htm

After fully updating Windows 10 Pro 32bit so that all the Powershell files are updated, this was my result:

System Meltdown Protection: YES
System Spectre protection: NO!
Performance: Good

Again, Windows is up to date - But it does say system is not immune to Spectre.

App was updated 27-1-2018 acording to the website by Gibson Research.

Then I installed the Ashampoo Meltdown Spectre CPU tester:

Spectre gives the red alert.

Meltdown says: Secure

After I found your page above, you say Atom N series should not need a BIOS update because the function is not in the CPU - probably the side band addressing.

But these tools report that the N270 of the Asus 1000H Netbook is not immune.

What could I do the really find out this nebook is immune or not immune, because I think I will not get an CPU firmware update for this system.

MS says it will give later updates for the 32bit version of Windows 10 as this cpu is only 32bit Atom, but I think this only is for meltdown.

According to your list on this page you say this cpu, this atom n - should not be effected.

Could you let me know how I really found out if I can do safe banking on this netbook and if this atom cpu misses the brance prediction instruction set, so that it will be immune.

No idea if I can trust these 2 apps.

Hope you can help me.

Have a great day,

Regards!


#45

I will get back to you on this tomorrow evening (GMT).

The best test is to simply run the Spectre Proof of Concept Code.
I’m sure in the meantime someone else can link to a safe open source version from github for windows.


#46

I’m going to make a thread on using hardware for a week that is spec-exec-theft immune. I have a SPARC box and a netbook with an ATOM N270.


#47

Well, time to get the ol’ Sharp Mebius out of the closet…

blows off dust

image

Would be interested to see an ARM-based laptop that can take a fair bit thrown at it. Maybe ARM and/or Qualcomm can take this opportunity. I know Qualcomm has already started out with the SD835, but the windows performance on those leaves a fair bit to be desired. Might be good for Android/Linux, though.


#48

What about Novena DIY laptops? Looks like they run A9 processors.


#49

What in the petty fuck is that and how do I get 10 of them.


#50

Read the post?

Unless you are a 4-year-old girl, trust me when I say this, your hands are too big for that laptop’s keyboard. As for how to get 10 of them… you’re going to have to scour pretty hard for them, as far as I’m aware they were never sold in the US.

Get a Pitop or a Pinebook. These days a A53 processor with a fast enough clock is probably much faster. These things were a neat idea in 2003 that could have been amazing if Intel didn’t get their panties in a twist over VLIW x86 emulation - the very reason that Transmeta processors are completely immune to the vulnerabilities.

This is much-needed karma for what Intel did to them.


#51

Thanks for the reply - I could not find a way to test this more or find safe code to test on github.

I might later just install Windows 7 on it again and run some other tests, but I do not know to make sure the CPU is immune.

I think Atom N series is not immune to spectre.

If you did find a document or test saying otherwise, then for now this should be considered not immune.

Regards,
M


#52

Can EPYC server CPU’s be considered safe?

I’m waiting until I graduate and have some real cash coming in to cycle my e5-2670 enclave out of commission and replace them with more secure and energy efficient alternatives.

In the mean time are there any ARM servers / motherboards available (not raspi) that aren’t vulnerable to spectre?

:evergreen_tree::goat: