Question Two different SSDs perfomring poorly

Jan 26, 2020
11
0
10
I've had two different SSDs now (both Sandisk, one 64gb and one 500gb) that perform well below expectations. I'm not getting a noticeable boot time difference and according to CrystalDiskMark my read/write speeds are indeed much lower than they should (~330/300). This is the second SSD now, I've tried swapping SATA ports on my MOBO (both SATA3), made sure AHCI was enabled in BIOS, checked for new FW, and optimized the drives in windows.

HW if helpful:
AMD FX-6300
ASRock 960GM/U3S3
G.Skill RipJaws X Series 8Gb RAM
Radeon R9 280x
EVGA 500W Power Supply

So unless the two different cables I've tried or both somehow faulty I'm not sure what could be wrong here...Any thoughts? Thanks in advance.
 
2 x SATA3 6.0 Gb/s connectors by ASMedia ASM1061, support NCQ, AHCI and Hot Plug functions
- 4 x SATA2 3.0 Gb/s connectors, support RAID (RAID 0, RAID 1, RAID 10 and JBOD), NCQ, AHCI and Hot Plug functions

Which one are you using?

SATA3 3.0Gb/s limit is around 300MB/s. if you're using the 6.0Gb/s you should see better speed indeed.
 
  • Like
Reactions: Maxxify
Jan 26, 2020
11
0
10
Hmm where did you find that info from? Way back when I first installed the card my research indicated 500W was just enough. Looking again I can't seem to find official documentation yet but I'm seeing some posts about 750W being recommended for crossfiring two of them.
 
I've tried both of my SATA 3 6.0 Gb/s connectors on my motherboard as shown in Maxxify's post. Looking at CrystalDiskInfo it is indeed set at SATA/600 | SATA/600.
Shown here.

That looks correct. You can update the driver for the ASMedia storage controller, I uploaded a set here. Please make a system restore point before doing so, just in case. It's also possible to update the controller's firmware but due to risk I would not suggest doing so directly - updating your motherboard's BIOS to 1.50 instead, if you haven't, may update the related OpROM.

I actually have an adapter based on the ASM1061 so I'm familiar with its workings. It's not super performant in comparison to other 6G solutions, but it should be above 300 MB/s.
 
Jan 26, 2020
11
0
10
Updated the driver with the link you provided and updated the BIOS to 1.50 following ASRock's documentation but I'm afraid nothing has noticeably changed. Reads/Writes are still struggling to get much higher than ~300 MB/s.
 
Not sure what else to check. You stated you set BIOS to AHCI - make sure this is for the "Onboard SATA3 Operation Mode" as that is the one that applies to the ASM1061 ports (the other one, "SATA Operation Mode," is for the 3G ports). The ASM1061 should be the limiting factor so I don't think other drivers will help much, e.g. chipset drivers, although if you want to check I uploaded Snappy Driver Installer here. You want to extract its files to a folder, run the _x64 .exe, check the "Create a restore point" box, and choose "Download Indexes Only" on the Welcome screen. This will download and check for new drivers system-wide. It's possible something was missed in your updates, if so it'd probably be related to AMD/chipset drivers, beyond that I have to think there's something interfering with the drive on the software end.

Windows has something called Resource Monitor (Start menu) that you can use to track the active time of the disk. If it's hitting 100% during CrystalDiskMark, for example, then there might be a software issue. This is pretty much the extent of my advice that doesn't require more heavy lifting.
 
Jan 26, 2020
11
0
10
Double checked BIOS, both "Onboard SATA 3 Operation Mode" and "SATA Operation Mode" are set to AHCI. Also tried Snappy Driver and did the missing drivers and still the same result.

I ran CrystalDiscMark with the Resource Monitor, it was hitting 100% the entire time.
 
Double checked BIOS, both "Onboard SATA 3 Operation Mode" and "SATA Operation Mode" are set to AHCI. Also tried Snappy Driver and did the missing drivers and still the same result.

I ran CrystalDiscMark with the Resource Monitor, it was hitting 100% the entire time.

Yeah, if the drive is listed as 100% duty cycle (active 100%) that means there's likely overhead somewhere, if that makes sense. In other words, the performance it's showing is 100% of its capability, if the results are lower than expected then something is siphoning off the bandwidth, e.g. a backup agent, anti-virus, driver verifier, whatever. Or it might simply be the maximum capability of the ASM1061 with the loss going to overhead - although I have some ASM1061s myself and they do get in the 400+ range anyway. You can rule out drivers/software by testing in Safe Mode.
 
Jan 26, 2020
11
0
10
Ok I'll try running in safe mode to try eliminating more possibilities. However this is all being done on a fresh install of Win 10 already so there's not much there to cause conflicts.

I also had the 500 gb SSD tested on a friend's computer. CrystalDiscMark speeds were showing up as ~500 MB/s so there's definitely something wrong here...
 
I also was disappointed at one point, when I bought a Samsung Evo SSD. I has an AMD FX-4300, and was disappointed by the very small increase in loading and boot times. I tried all the motherboard ports, and none got any better. I thought the SSD was overhyped. Over a year later, I upgraded to a Ryzen CPU, with a whole new build. I kept the SSD. It was immediately much faster. I was actually shocked by how much the FX series CPU and the board I had was limiting that SSD. The boards appeared to be rated very similarly for speeds and everything, but I'm guessing a possible limitation with the FX CPUs themselves could be an issue here. Just thought it might be interesting, since I had a similar experience myself. Though I never did any speed testing on it myself.
 
Jan 26, 2020
11
0
10
Update: Still no change after bench-marking with CrystalMark in safe mode. Seems as though there's truly something fundamentally wrong with my build that is causing the issue. Whether its the MOBO, CPU or something else I have no idea.

I was intending on building a new PC in the near future so I suppose this is just another reason to do so.