Phison E18 slow writes certain file sizes on Atto

[mechtech]

Just wondering if this is a FW issue or an atto issue or other?? It carries through my KC3000 1TB & 2TB and Corsair 4TB drives.

Just curious is anyone else gets those anomalies with phison E18 drives?? @W1zzard

 

[W1zzard]

I'm not using ATTO in my testing, any chance your SLC cache is full at that time?

 

[mechtech]

I'm not using ATTO in my testing, any chance your SLC cache is full at that time?

Hmmmm

not sure. Tested on a fresh start up and after I did a trim on the drive.

 

[W1zzard]

How much free disk space?

 

[mechtech]

How much free disk space?

90% free

edit - just wondering if anyone else can test if they have an nvme with a Phison E18 controller??

 

[K²mix]

Looks more like something is affecting the Atto benchmark?
I don't see any abnormalities with mine. Performance comparable to new condition. No KC3000 but Phison E18 +112 layer BICS5.

 

[mechtech]

Looks more like something is affecting the Atto benchmark?
I don't see any abnormalities with mine. Performance comparable to new condition. No KC3000 but Phison E18 +112 layer BICS5.

View attachment 369589

Thanks

MS nvme driver? (that's what I'm using, and amd chipset driver)

 

[Harlow]

I've just run ATTO on my KINGSTON SKC3000S1024G and the results are almost the same as those in K2mix's chart, but reads are a bit faster at 8.65MB/s.

I have see inconsistent results with some drives in ATTO, but have put it down to cache hits or controller latency at some test file sizes and stopped worrying about it.

 

[egg_addict]

It seems that drives with a Phison E18 controller have issues with write speed degradation over time. I believe Kingston has a firmware update that fixes the issue, not sure what other manufacturers have an updated firmware that fixes the issue. If you've had these drives for a while, that's probably why.

 

[massaker]

It seems that drives with a Phison E18 controller have issues with write speed degradation over time

You mean READ speed degradation... Yes, it has.

I believe Kingston has a firmware update that fixes the issue, not sure what other manufacturers have an updated firmware that fixes the issue.

Have already read (some user experiences) it didn't really helped much.

If you've had

It would not affect benchmarks with fresh written data...

 

[Space Lynx]

I've just run ATTO on my KINGSTON SKC3000S1024G and the results are almost the same as those in K2mix's chart, but reads are a bit faster at 8.65MB/s.

I have see inconsistent results with some drives in ATTO, but have put it down to cache hits or controller latency at some test file sizes and stopped worrying about it.

there is a firmware update for kingston kc3000, i was actually looking at updating mine earlier today

august 2024 is the release date for the firmware

SSD Firmware Update - Kingston Technology

Kingston SSD firmware updates are provided via Kingston’s SSD Manager (KSM) application. Always remember to backup your data before proceeding with a firmware update.

www.kingston.com