Write caching is using the RAM instead of hitting the NAND immediately. Only turn write caching on if you have a backup power supply, like a UPS. Otherwise, you risk losing data.
Also, I'm beginning to think that this SSD has a really crappy controller or it is using QLC. Either way, I would stop trying to "fix" it and just get a better SSD. That's the problem with saving a few dollars when it comes to PC's. You may tihnk that you did the smart thing but it just ends up making your life worse.
Also, I'm beginning to think that this SSD has a really crappy controller or it is using QLC. Either way, I would stop trying to "fix" it and just get a better SSD. That's the problem with saving a few dollars when it comes to PC's. You may tihnk that you did the smart thing but it just ends up making your life worse.
Write-Caching is enabled by default Win10 for system drive as far as i can tell.
I have tested new Crucial MX500 500GB (has 512MB DRAM) over SATA with Windows Write-caching disabled and i get same low write speeds as DRAM-less Lexar.
P.S. I am not worried too much about data loss on SSD in case power is off, since i use SSD only for OS, all my data is stored on external USB IDE.
PNY CS3030 1TB - $110 / Phison Controller E12
BPX PRO 1TB - $130 / Phison Controller E12
So, comparing essentially the same drives the #'s vary considerably depending on the performance you're looking for non either R or WR.
CS3030 is a better reader
BPX Pro is a better writer
Both use the same controller so, not an issue when looking for apples to apples. It comes down to other factors such as the space in use / ratio, which slot they're in potentially, OS drive or not, etc.
I have a pair of SN850's in my server but, CDM doesn't work with Linux and from a CLI test using DD though they hit advertised speeds w/ Gen 4 controllers ~2X these shown above.
Unfortunately there's no clear cut reason why 2 drives with the same controller perform differently when it comes down to the components being very similar and IIRC they even use the same PCB to attach them to.
Either way w/ these 2 NVME's I can hit 1.5GB/s moving files between them in real world moves not just the synthetic testing.
BPX PRO 1TB - $130 / Phison Controller E12
So, comparing essentially the same drives the #'s vary considerably depending on the performance you're looking for non either R or WR.
CS3030 is a better reader
BPX Pro is a better writer
Both use the same controller so, not an issue when looking for apples to apples. It comes down to other factors such as the space in use / ratio, which slot they're in potentially, OS drive or not, etc.
I have a pair of SN850's in my server but, CDM doesn't work with Linux and from a CLI test using DD though they hit advertised speeds w/ Gen 4 controllers ~2X these shown above.
Unfortunately there's no clear cut reason why 2 drives with the same controller perform differently when it comes down to the components being very similar and IIRC they even use the same PCB to attach them to.
Either way w/ these 2 NVME's I can hit 1.5GB/s moving files between them in real world moves not just the synthetic testing.
KDiskMark Is A GUI HDD / SSD Benchmark Tool For Linux (Similar To CrystalDiskMark)
KDiskMark is a free and open source alternative to CrystalDiskMark (which is Windows only) for Linux, a GUI HDD / SSD benchmarking software.
www.linuxuprising.com
Somewhat similar.
Meh, its a headless setup and not worth the hassle for testing. CLI / DD testing already shows good speeds.KDiskMark Is A GUI HDD / SSD Benchmark Tool For Linux (Similar To CrystalDiskMark)
KDiskMark is a free and open source alternative to CrystalDiskMark (which is Windows only) for Linux, a GUI HDD / SSD benchmarking software.
Somewhat similar.
IntelUser2000
Elite Member
- Oct 14, 2003
- 8,686
- 3,786
- 136
It's normal. TRIM takes very short time to run unlike defragmentation. TRIM works with the internal algorithm and gives a go to let cells be deleted so it can be ready for use in the future, rather than having it kick in as it's needed.
@Tech Junky Firmware differences account for the rest. There's likely bunch of knobs that can be adjusted at a firmware level and it'll be a balance of lot of different factors such as sequential vs random, read vs write, burst vs sustained, high performance versus lower power.
Last edited:
I am seriously frustrated about this whole SSD thing, its regarding Crucial MX500 500GB (500MB DRAM)
Using Crucial Storage Executive software's Momentum Cache feature, it uses up to 4GB of ram to cache the HDD operations.
In theory its okay however when i do work with copying many several gigs of files i get into problem.
When i copy this folder in about 5 seconds, however then in one minute i am getting to the point that my whole system freezes because SSD has 100% utilization and anything i try to open from SSD fails at that time and my OS obviously also freezes until copy finished.
SSD supposed to be used 100% when writing?
Using Crucial Storage Executive software's Momentum Cache feature, it uses up to 4GB of ram to cache the HDD operations.
In theory its okay however when i do work with copying many several gigs of files i get into problem.
When i copy this folder in about 5 seconds, however then in one minute i am getting to the point that my whole system freezes because SSD has 100% utilization and anything i try to open from SSD fails at that time and my OS obviously also freezes until copy finished.
SSD supposed to be used 100% when writing?
At this point you're just getting frustrated and it's time to drill some holes in it and get something that works.
I've used Crucial RAM in the past and not had an issue but, never used their HDD/SSD's as well, they weren't impressive enough to consider.
For a 2.5" 1TB SATA just get a Samsung and call it a day. I've got one sitting here in a box as a spare ICE drive that I've had since ~2015 and never an issue. Other than expense $$$$ back then. Couple of other ones sitting in another box at 128/256GB sizes. Even have a OCZ 32GB sitting around.
I've used Crucial RAM in the past and not had an issue but, never used their HDD/SSD's as well, they weren't impressive enough to consider.
For a 2.5" 1TB SATA just get a Samsung and call it a day. I've got one sitting here in a box as a spare ICE drive that I've had since ~2015 and never an issue. Other than expense $$$$ back then. Couple of other ones sitting in another box at 128/256GB sizes. Even have a OCZ 32GB sitting around.
Oh, hello, fellow OCZ user! Mine was the OCZ Vertex 30GB, my first ever SSD and also one that reached end of life (5% life remaining) in about a year.
Problem is fixed so far after i enabled write caching in windows 10, originally it was on, but i turned it off when i enabled Momentum Cache, now i disabled Momentum Cache. And load on HDD goes up to 80% sometimes 99% very briefly, i don't experience any lag.
How i increase SATA voltage, perhaps i have old PSU its 15 years old or so, but had very light use.
mikeymikec
Lifer
- May 19, 2011
- 20,007
- 14,344
- 136
Wouldn't hurt to check the SMART diagnostic data on the SSD, post it here if you need help interpreting it.
mikeymikec
Lifer
- May 19, 2011
- 20,007
- 14,344
- 136
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 20K
-
-
-
Facebook
Twitter