6 Replies Latest reply on Mar 29, 2018 1:49 PM by N.Scott.Pearson

    Should I expect any solutions after the bugged Spectre & Meltdown updates?

    Namikage

      Hello Intel,

       

      I'm only doing this after the unsatisfying feedback am receiving regarding the obvious slowdowns caused by the bugged Spectre & Meltdown patches. What I understand is that these updates are firmware updates that are stored in the ROM. They are not the usual updates windows receive in the sense there is no way to uninstall them. The only way for them to be overwritten is to get a new firmware update to resolve the issue. My issue concerns the slowdowns caused by these patches. These slowdowns are seen in the delayed system sounds and the freezing of my video player due to delayed sound reading. Pretty much, everything takes a bit more time and the slowdowns are real. What I need to know: Would there ever be a firmware update that will fix this for me? These are some information you may need: Processor Name: Intel(R) Celeron(R) CPU 877 @ 1.40GHz, Version: 5.80.20171018, Windows 7 64bits. I have only the meltdown patch installed because I was using Iobit Driver Booster to instal the patches and it got stopped in the middle of updating. I used InSpectre.exe to make sure. Hope you provide me with something.

        • 1. Re: Should I expect any solutions after the bugged Spectre & Meltdown updates?
          Intel Corporation
          This message was posted on behalf of Intel Corporation

          Hello Namikage,

          We appreciate that you are taking time to contact Intel about your concerns. We understand that the security and performance of your devices are important to you. Performance impacts depend on workload and configuration. For Intel’s latest performance testing results, see https://newsroom.intel.com/editorials/intel-security-issue-update-initial-performance-data-results-client-systems/ 

          Please check with you operating system vendors and system manufacturers, and apply any available updates as soon as they are available. You should do this whether you use an Intel-based system, or other computer or mobile device. The following link provides full details on this topic: https://www.intel.com/content/www/us/en/architecture-and-technology/facts-about-side-channel-analysis-and-intel-products.html 


          Regards,
          David V

          • 2. Re: Should I expect any solutions after the bugged Spectre & Meltdown updates?
            Namikage

            Hello David V from Intel,

             

            Unfortunately, you didn't answer my question. I was hoping you say something like: Yes, we will be providing firmware updates that will restore your CPU to its optimal performance before Your partial update of the the bugged Spectre & Meltdown patches. I was hoping you tell that I have nothing to worry about the fact I'm gonna be using a faulty CPU due to a buggy firmware update. I'm simply asking you Intel to at least provide me with a way to restore my firmware as it was. I don't care about this Spectre & Meltdown vulnerability. I just want these annoying Delays, lags and slowdowns to stop. If you cannot fix these vulnerabilities fully without causing any issues to us PC owners, please give us a way to restore my firmware as it was. You are Intel after all. Thanks.

            • 3. Re: Should I expect any solutions after the bugged Spectre & Meltdown updates?
              Intel Corporation
              This message was posted on behalf of Intel Corporation

              Hello Namikage,

              Thank you for your response,

              On performance, Intel continues to conduct testing, develop ways to improve performance, and provide updates to minimize performance impacts. Our testing results to date show performance impact that ranges depending on specific workloads and configurations. Our engineers are working to address identified performance effects. 


              Regards,
              David V

              1 of 1 people found this helpful
              • 4. Re: Should I expect any solutions after the bugged Spectre & Meltdown updates?
                N.Scott.Pearson

                Let's make sure that this is understood properly...

                 

                There are essentially three vulnerabilities to be discussed, SpectreA, SpectreB and Meltdown. The microcode updates being released by Intel only address the SpectreB vulnerability. I don't know for sure, but I don't believe that these microcode updates cause any slowdown in processor operation. The SpectreA and Meltdown vulnerabilities, on the other hand, cannot be addressed in microcode; changes in the silicon design will be required to address them. This can only happen in new processor designs. Bottom line, you will never see a microcode update that addresses all of these vulnerabilities.

                 

                For existing processors, workarounds in the O/S are necessary to address the SpectreA and Meltdown vulnerabilities. It is the nature of these workarounds that is responsible for the slowing of the processor. For example, when making a system call (i.e. transitioning from User (Ring3) to Kernel (Ring0) privilege level and back again), data must be copied back and forth to ensure that privileged data is not leaked. It is the cost (overhead) of these copy operations that slows the processor. As Intel has indicated, in typical application designs making typical numbers of system calls, the cost of these copy operations is small (something like 7%). When applications make higher numbers of system calls, the slowing can be worse.

                 

                I speculate (pun intended) that most of the issues that you have been seeing are not typical. I believe that they are the result of the early microcode releases, which contained bugs that caused system stability issues. If you get and install (once available) an updated firmware package from your board manufacturer that contains the updated microcode needed for your processor, you should not see these stability issues any longer. As for the workarounds in the O/S, if you don't like the slowdowns incurred and want to disable them (something I do not recommend as you will then have a vulnerable system), you need to talk to your O/S vendor (Microsoft for Windows, Distribution owners for Linux) regarding whether they provide a way to disable the use of the workarounds (I have heard conflicting stories on how you can do this in Windows).

                 

                Hope this helps,

                ...S

                1 of 1 people found this helpful
                • 5. Re: Should I expect any solutions after the bugged Spectre & Meltdown updates?
                  Namikage

                  Thank you for your replies.

                   

                  First, I'm not asking for a microcode update that resolves these vulnerabilities. I'm asking for a solution regarding the CPU performance. That alone is easy because I don't think it is impossible for them to write something that will return everything as it was before these updates. That is for now the fastest solution possible. Workarounds through software are way safer than messing with the firmware. The overall issue here is that Spectre & Meltdown vulnerabilities were mitigated through disabling CPU's Speculative Execution. This literally helps the CPU to speculate all the routes any software my require before it even requires it. This means, at this moment my PC CPU is working without that feature. That is why, there are 2% to 10% slowdowns, of course, depending on you PC hardware. Now, David V ensured that they are working on a solution to address identified performance effects. For me as a PC user, it is enough. Thank you.

                  • 6. Re: Should I expect any solutions after the bugged Spectre & Meltdown updates?
                    N.Scott.Pearson

                    The slowdowns being seen have absolutely nothing to do with the state of the CPU's Speculative Execution engine. It is the workarounds that were introduced into the O/Ss to handle privilege state transitions during O/S system calls that are responsible for the slowdown being seen.

                     

                    Let me be clear: These slowdowns cannot be eliminated. It is an increase in the overhead for making O/S system calls. It will be incurred by all existing processors. What is being investigated is whether there are alternative ways to implement these workarounds that will result in a reduction in the amount of overhead incurred. I am not optimistic that alternative(s) will be identified.

                     

                    ...S

                    1 of 1 people found this helpful