Troubleshooting • Re: Issues with USB 3.0 vs 2.0 Raspi 4B

64GB kingston DT

Run 1
prepare-file;0;0;71800;140
seq-write;0;0;65552;16
seq-read;86946;21;0;0
rand-4k-write;0;0;207;51
rand-4k-read;4188;1047;0;0
Sequential write speed 65552 KB/sec (target 10000) - PASS
Random write speed 51 IOPS (target 500) - FAIL
Random read speed 1047 IOPS (target 1500) - FAIL
Run 2
prepare-file;0;0;15791;30
seq-write;0;0;17823;4
seq-read;87997;21;0;0
rand-4k-write;0;0;5;1
rand-4k-read;6201;1550;0;0
Sequential write speed 17823 KB/sec (target 10000) - PASS
Random write speed 1 IOPS (target 500) - FAIL
Random read speed 1550 IOPS (target 1500) - PASS
Run 3
prepare-file;0;0;15397;30
seq-write;0;0;18685;4
seq-read;86601;21;0;0
rand-4k-write;0;0;9;2
rand-4k-read;5293;1323;0;0
Sequential write speed 18685 KB/sec (target 10000) - PASS
Random write speed 2 IOPS (target 500) - FAIL
Random read speed 1323 IOPS (target 1500) - FAIL


1-2 random IOPS? Yet it seems better than the other one...

Plugging them into the front panel of my USB can clearly measure the 128gb drive out performing the 64gb in crystaldiskbench for 4k random IOPS as well (not sure if that is the same as the diagnostic). Either way it shows a consistent increase as expected perf wise where as the diagnostic tool + a DD write of 1gb shows inverted behavior.

While the problem is device specific (probably the controller used between the two) I believe the root cause would be driver related as a guess. Power is not throttling, singular devices at 504 should not be impairing performance. The throughput should be the same and in windows the performance is consistent with the newer generational USB.

Statistics: Posted by solace50 — Tue Sep 09, 2025 11:10 pm

---