Question Crucial® BX500 1TB Randomly Slow Read Speeds

Status
Not open for further replies.

fedefrasis

Distinguished
Dec 16, 2014
113
0
18,690
Hello. This are my specs
I9 9990KF
Asus Maximus Hero
64Gb Ram
2xWD Blue SDD Sata
2xCrucial BX400 SSD Sata

Both crucial drives were bought together. At install they ran perfectly. Then they started slowing down to like 90mbps. Then the problem seemed to go away...
The crucial drives randomly go down to like 90mbps reads. Then climb back to 190ish, then down to 90ish again. Some days they just work flawlessly and run at their full speed. This problem had gone away for a long period and now returned. Does someone know what causes these fluctuations? And no, the drives being emptier doesnt change it much. I have seen them slow, while at 30% capacity and running at full speed at 80% So i really dont know. Also read speeds seem to be impacted more heavily

One is at 30% capacity the other at 70. This is after running right click->optimize from This Computer

What can I do about this? Thanks!
 
I know exactly what you're talking about and the BX500 is notorious for this issue. If yours is ONLY dropping to 90MB/s, it's doing way better than mine. Sequential reads on mine can drop to ~5MB/s. I suspect it's data that hasn't recently been written that's most affected. The issue doesn't tend to show up clearly in CrystalDiskMark, probably because it writes a new/fresh test file and it goes with peak readings. If you run a full drive READ scan, using something like HDDScan or Victoria, I'll bet the transfer rate graph shows some wild swings.

I have tried just about everything and the only (temporary) solution I've found is to wipe the drive and rewrite the data (even if it's the same exact data). This restores the performance but it will slowly degrade yet again. I'd love to know what's going on. And, I don't think it's just because this drive is DRAM-less. I've seen plenty of DRAM-less drives that do not behave this way, and I've seen a pretty decent NVMe drive, with TLC NAND and the ideal amount of DRAM, that does do this.

Would you mind sharing screenshots of their CrystalDiskInfo SMART attributes and at least the first four digits of their serial numbers?
 
View: https://imgur.com/a/aNyDaa0

I know exactly what you're talking about and the BX500 is notorious for this issue. If yours is ONLY dropping to 90MB/s, it's doing way better than mine. Sequential reads on mine can drop to ~5MB/s. I suspect it's data that hasn't recently been written that's most affected. The issue doesn't tend to show up clearly in CrystalDiskMark, probably because it writes a new/fresh test file and it goes with peak readings. If you run a full drive READ scan, using something like HDDScan or Victoria, I'll bet the transfer rate graph shows some wild swings.

I have tried just about everything and the only (temporary) solution I've found is to wipe the drive and rewrite the data (even if it's the same exact data). This restores the performance but it will slowly degrade yet again. I'd love to know what's going on. And, I don't think it's just because this drive is DRAM-less. I've seen plenty of DRAM-less drives that do not behave this way, and I've seen a pretty decent NVMe drive, with TLC NAND and the ideal amount of DRAM, that does do this.

Would you mind sharing screenshots of their CrystalDiskInfo SMART attributes and at least the first four digits of their serial numbers?
 

In CrystalDiskInfo, could you go to Function > Advanced Feature > Raw Values > 10 [DEC]. That will make things more readable to us meatbags. Then, could you scroll to the bottom and take another screenshot? I'd like to see the last two attributes, which can be used to calculate the write amplification.

From what I can see, nothing jumps out as abnormal. Then again, nothing looks weird in the SMART attributes on mine either. Yours is quite a bit newer and has a much higher firmware version. It makes me wonder how many hardware variants there are of this drive. If the MX500 were anything to go by, it could indicate that yours is five hardware revisions newer than mine. However, I'm not at all confident in that assumption.
 
Status
Not open for further replies.