Here is an AIDA64 Engineer System Stability Test:
Here is an AIDA64 Engineer System Stability Test:
Hi everyone, I have some questions about whether my AIDA64 Engineer system stability test worked well. Here are my computer specs: an AMD Ryzen 9 3900X Processor (12 Cores/24 Threads) running on a custom-built machine built by a local IT company in Melbourne, Australia. The PC has a Motherboard that includes WiFi and Bluetooth, plus RAM from Team T-Force Vulcan (64 GB DDR4 3200 MHz). I also have an NVMe Solid State Drive from Gigabyte, a Graphics Card from Nvidia GeForce GTX-1050Ti (4 GB), and a case with a power supply. Michael, the guy who built my PC for me, suggested running this stability test using AIDA64 Engineer. He wrote in an email that I should pick "Stress GPU(s)" on the top left and press start in the bottom left. That window lets you check system temperatures and CPU usage, or even if the CPU gets throttled. Later, Michael explained what the test does: it makes a fake load to stress the CPU, motherboard, RAM, GPU, and drive(s). These parts are key places things can break. I totally forgot about running this stress test for months until yesterday. Michael has since left my IT company, so I contacted them again to talk with Ryan G., who told me he doesn't know AIDA64. Since then, I got confused about whether to check the "stress local disks" box before starting. After re-reading Michael's email which said the test stresses the CPU, motherboard, RAM, GPU and drives, I decided to check that box instead. Then I ran the stability test for three hours and thirty-two minutes with these boxes checked: Stress CPU, Stress FPU (floating point), Stress cache, Stress system memory, Stress local disks, Stress GPU(s). Here are my results from the first run. I stopped early because a message board thread told me to avoid write tests on SSDs since they wear out the flash memory cells over time. So I ran the test again for one hour and eleven minutes with five boxes checked: Stress CPU, Stress FPU, Stress cache, Stress system memory, and Stress GPU(s). Here are my results from that second run. I have three questions now. Are my first test results good? Is my second test results good? How much did I wear out the flash memory cells in my 1 TB PCIe Gen4 Gigabyte NVMe SSD during those long runs with local disk stress? If I want to do a future stability test, which boxes should I check and how long should I run it for? And finally, how do I read the AIDA64 results, and what counts as a good result?
Aida64 is really good at checking how stressed a computer gets and it costs nothing to run. I've been using this tool with all my game builds lately. The numbers show that the CPU is getting too hot, almost always around 90 degrees Celsius. That means your cooling system isn't working well enough. You should try keeping things under 80C while playing games instead of letting them go up higher. Tell us what kind of cooling setup you have so we can see if it's enough. The third number is hard to read, but I think the SSD will probably be fine even with some wear on it from normal use. According to the Aida64 forum support team, this test makes all your hard drives work at the same time, only as long as they have a partition table and actually have files stored in them. This happens because the test is meant to check how well the file system works, so it needs those files to run. They say you shouldn't use the write tests on SSDs because that can make the flash memory cells inside get worn out over time. You...
Aida64 is a great stress tester and it's free. I use it with all my builds. 1: Both tests show high CPU temps around 90C, so your cooling system isn't working well. You should aim for no more than 80C under load. List out your cooling system? 3: It's hard to tell the wear on your SSD from this test alone but I doubt it will be much. On the Aida64 forum, the System Stability Test checks all local disks at once, but only if they have a partition table and files on them. This is because that test looks at the file system, so it needs files to run. They say don't use write tests on SSDs because they wear out flash memory cells quickly. You can still use the new Linear Write + Verify test from AIDA64 Disk Benchmark once on a new drive to check data integrity. That only uses one write cycle per cell, so it won't hurt much. 4: Run the initial stress test with CPU, FPU, and Cache checked off, then run other tests separately. Start by testing for 10 minutes and stop if temps go over 80C. Once you're sure everything is stable, run the full stress test for one hour. Use AIDA64 together with HWinfo64 to check temperature readings on the chip when it's under load, and check your RAM rail voltages to make sure they don't spike outside of 5% or more. I don't use the RAM test but for checking DIMMs, I use Memtest86 from a USB stick.
Hi, thanks a bunch for your nice message. I just want to clear up exactly how I should run AIDA64 going forward. It sounds like you mean this: Run the AIDA64 stress test with only the check boxes for CPU, FPU and Cache turned on for 10 minutes. See if the CPU temperature goes over 80 degrees Celsius on that first run. If it doesn't go over 80 degrees on the first test, then I should run the same test again for one hour. See if the CPU temperature goes over 80 degrees Celsius on the second run. Is my guess above right? What did you mean when you said "other tests run separately"? Should I use the "Stress GPU(s)" test? If so, how do I do that? What should I look for in the results? Should I use the "Stress system memory" test? If so, how do I do that? What should I look for in the results?
Yes, that setup works as written. When testing the CPU, FPU, and Cache, we are mainly checking how hot they get. I don't run all these tests at once; I do them one by one instead. There are more stressful GPU tests than AIDA64 has to offer, so I use games like Heaven or Cinebench for that. Aida64 is just one tool in your toolbox, and it's not complete on its own without adding other stress tests. AIDA64 with Just Memory works well if you want to see how stable a memory overclock is by itself, or when combined with other tasks. The Memory Latency benchmark checks how slow the CPU feels when it needs to read data from your system RAM. That "latency time" means how long it takes after you tell the computer to get data until that data actually arrives at the processor's registers. I think Realbench is the best real-world tester, and if you pass that test, your whole system is 100% stable.