3

My computer is old, but still powerful, yet I have a problem where things are "just slow" regularly applications all freeze. Most noticeably, Chrome takes a measurable amount of time to update after input, but it happens to other programs too (I know Chrome is often resource hungry, but I don't seem to be hitting resource limits). For example, as I am typing this question, I tried to use the arrow keys on the keyboard to move back in the textbox to correct a typo and it takes a noticeable amount of time from when I press the arrow key until the cursor actually moves, so I often end up moving too far. Additionally, I can often type a full sentence before any of it will appear in the textbox. Every so often, every program will just stop responding completely, usually not more than 30 seconds, but sometimes for up to 10 minutes.

Spec:

  • Dell Precision Tower T7500
  • 2x Intel Xeon X5550 (4c/8t, 2.66GHz, seen boosting to 2.9GHz, listed to 3.06GHz)
  • 48 GB RAM (12x Kingston 9965447-001.A00LF 4GB 1RX8 DDR3-1333 PC3-10600 ECC Registered RDIMM)
  • OS Disk: 2 TB RAID 1 (Dell 6/iR RAID controller) - I think Western Digital Blue disks
  • Additional Disks: 256 GB Samsung SSD, 2x 500 GB Seagate HDDs
  • Graphics Cards:
    • Original: 2x ASUS GT 710
    • Current: EVGA GTX 660 Superclocked, Quadro FX 3800
    • Add-in USB3 card
    • Windows 10 Education Edition

I have been watching the Performance tab in Task Manager and it doesn't seem like any specific resource is hitting limits. As I type this (with the issues mentioned above):

  • CPU load: 15%-20% total, no thread shown over 75%
  • RAM usage: 19.6/48 GB (even with 100 Chrome tabs open)
  • Disk active time: 20% - 50%
  • GPU Utilization: 0% - 5%

I have had this problem for a long time, but now I am working at home and can't get anything done. Things I have tried recently:

  1. Checked all drivers and firmware are up to date.
  2. Moved pagefile to separate HDD.
  3. Switched graphics cards from passively cooled to actively cooled (cards were running hot according to CoreTemp).
  4. Moved Chrome cache directories to SSD.

None of these have had a substantial impact on the problem. What else should I check?

Improve this question
6
  • Try : (1) Check SMART info of the hard disk and SSD (if supported). (2) Check the Event Viewer for alarming messages. (3) Do sfc /scannow.
    – harrymc
    Commented Mar 25, 2020 at 14:27
  • SMART Info: RAID card reports drives are healthy. CrystalDiskInfo reports one RAID drive has a warning on C6 Uncorrectable Sector Count (Current 200, Worst 199, Threshold 0, Raw: 11B) and the other has a warning on 05 Reallocated Sectors Count (Current 196, Worst 196, Threshold 140, Raw: 1F). All other drives are fine.
    – yakatz
    Commented Mar 25, 2020 at 14:33
  • 1
    WD Blue disks in RAID with warnings? Replace immediately, or data loss will almost certainly result. Also, retries while writing could be slowing you down since your drives aren't really meant for RAID. I can't find the article now, but I read years ago about how many consumer disks will transparently retry failed writes so the RAID controller doesn't know that the write actually failed.
    – Moshe Katz
    Commented Mar 25, 2020 at 15:00
  • CPU/Bus temperature could also be a factor. If your CPU is overheating (and Xeon ARE prone to overheating), it may be throttling, thus rendering your PC slow or sluggish. Try installing any monitoring tool that'll give you a direct, real-time temperature reading of your CPU. Speccy would do the trick for something so simple.
    – user1019780
    Commented Mar 25, 2020 at 15:44
  • @Didier Temperatures look fine. My old graphics cards were getting too hot, so I replaced them with actively cooled models.
    – yakatz
    Commented Mar 25, 2020 at 18:33

3 Answers 3

Reset to default
2

The S.M.A.R.T. attribute Uncorrectable Sector Count is a critical parameter and indicates the number of uncorrectable errors. The raw value of this attribute indicates the total number of uncorrectable errors when reading/writing a sector.

As disks contain replacement sectors for failed sectors, this may mean that this reserve was exhausted and bad sectors are no longer reallocated by the firmware.

Conclusion: The disk in question may be facing imminent drive failure. You should urgently do data backup and hardware replacement.

In RAID-2 this means that you need to backup, replace the disk, create a new RAID array and restore the data.

Improve this answer
1
  • You have a very old CPU/chipset/memory.
  • Single thread performance was always horrible in Core2Quad/first few gens of new Xeons. However, it was possible - with a few motherboards and of course with better cooling - to overclock these CPUs/motherboards. You may or may not be able to OC yours. On Youtube you can find nice vids/guides about this.
  • Since this PC was built, it was discovered that Intel CPUs have a truckload of vulnerabilities. A LOT. And I am not exaggerating. Note: Even if you disable hyper-threading the mitigations will still slow you down.

List of mitigations/bugs on an Intel i7 4770 (Linux /proc/cpuinfo):
bugs: cpu_meltdown spectre_v1 spectre_v2 spec_store_bypass l1tf mds swapgs itlb_multihit

List of mitigations/bugs on an Intel Xeon-D 1521 (Linux /proc/cpuinfo):
bugs: cpu_meltdown spectre_v1 spectre_v2 spec_store_bypass l1tf mds swapgs taa itlb_multihit

And these are way newer processors. Each one of these so called "mitigations" slow down your PC one way or another. Sometimes, significantly so.

So what can you do?

  • Buy a new computer. Going with the assumption you need many threads/cores (judging from the ram amount)...
    Intel still haven't released a new architecture - so it would soon become kinda obsolete. And a dual-socket Intel costs an arm and a leg. You could however pick up a first gen Threadripper, or a Ryzen and have many threads/cores. Even ECC memory.
  • Buy a used computer. You can pick up a used enterprise dual-socket AMD PC for example. AMD barely got hit by the mitigations/vulns, but the old architectures before were pretty horrible (ie.: Bulldozer).
  • Disable mitigations. Now this is a risky one. This is REALLY not recommended on a production server. But on a home PC, if you use an adblocker and have at least a regular anti-virus (even Win10 default works), then you should be good. This tool can help you disable some mitigations: https://www.grc.com/inspectre.htm
  • Windows also implemented a lot of security features that can further lead to slowdown, especially on older PCs. Go into PC Settings -> Update & Security -> Windows Security -> App & Browser control -> (go down in the list) Exploit Protection -> Exploit protection settings.

Screenshot of Exploit protection settings:

Screenshot of Exploit protection settings

Improve this answer
0

If that machine is lagging like what you are describing, you have some kind of a hardware failure going on. I have a similar generation machine and it is still very capable for most workloads.

Check your event log for machine check exceptions (MCEs), memory ECC/EDAC errors, do a full disk SMART self-test (short and long), check your cooling, and do a couple of hours of OCCT stability test.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .