Corsair Force MP600 PCIe 4.0 NVMe SSD Benchmarks On Linux

[andreano]

Something I would like to see in a storage device test (not so easy to automate, I admit) is to see if it will eat your data on power loss.

Like this guy is testing: https://www.youtube.com/watch?v=AHIld3cjmTM&t=11m13s

[kobblestown]

Yeah that's a known performance deficit for windows but it doesn't explain why the nvme on pci-e 4.0 would suck more on win compared to linux... that's what's troubling me.

For my money it's Spectre/Meltdown. In my observation Windows' IO has become signifficantly slower during the last year. My machine was able to delete big file trees with about 2000 files/sec for SATA SSDs and 3000 files/sec for NVMe SSD. Now it fluctuates big time and goes down to the lower tens (10-20-30) at times while never shooting above 1000. Tested on several different system. I have no other explanation for that. It makes no sense that an NVMe SSDs should be slower than an HDD.

So it boils down to (1) probably less efficient implementation of mitigations and (2) enabling more mitigations than necessary. Do you think Windows only enables the mitigations appropriate for the CPU, like disabling Meltdown and some of the Spectre mitigations when run on AMD CPUs? Well, that might actually be the case but I doubt it. Windows is shite! A steaming pile of.

Also, as far as I know, Windows requires more context switches for file I/O so the mitigations would have larger implact.

BTW, I'm not an authority on these issues.So take my post with a grain of salt. I would realy like to hear more from people whether Windows has become slower over the last year or so. I routinely delete deeply nested directories with hundreds of thousands of files in total in them. And the performance for that kind of operations has fallen off a cliff.

[_Alex_]

I used to work for a large enterprise SAN vendor, and Linux consistently outperformed Windows in storage benchmarks on identical hardware, sometimes by a quite a lot. The Windows storage subsystem just sucks.

Yeah that's a known performance deficit for windows but it doesn't explain why the nvme on pci-e 4.0 would suck more on win compared to linux... that's what's troubling me.

[torsionbar28]

Linux seems to perform faaar better.

I used to work for a large enterprise SAN vendor, and Linux consistently outperformed Windows in storage benchmarks on identical hardware, sometimes by a quite a lot. The Windows storage subsystem just sucks.

[_Alex_]

Interesting. Windows benchmarks showed nothing serious improvement-wise for this SSD... in many benchmarks it was lagging the old mp510...

Linux seems to perform faaar better.

And a correction: "The MP600 2TB costs around $450 USD while the Optane 900p 280GB costs $250 USD, but there is also the nearly four times greater storage capacity with the Corsair SSD." ....should write ~7 times greater storage.

[Space Heater]

I was able to upgrade the firmware on my MX200 with Linux.

I think you are confusing Crucial and Corsair. Crucial is one of the companies that does support bootable firmware update ISOs.

[Yndoendo]

Corsair does not provide any method to update their products' firmware without using Windows - they don't even provide an operating system agnostic firmware update iso.

Don't support companies that refuse to support your use case.

I was able to upgrade the firmware on my MX200 with Linux.
Just had turn the supplied ISO to be able to be booted with UEFI.

<h2>Unoffical Guide: Firmware Updating via UEFI with reFind</h2>
<p>&nbsp;</p>
<p>Say your usb flash drive is <code>/dev/sdg</code>; create a partition with your favorite tool:</p>
<pre>sudo cfdisk /dev/sdg</pre>
<p>Format the drive:</p>
<pre>mkfs.vfat -F 32 /dev/sdg1</pre>
<p>Create the boot system:</p>
<pre>sudo refind-install --usedefault /dev/sdg1 --alldrivers</pre>
<p>Mount the partition to file files:</p>
<pre>sudo mount /dev/sdg1 /mnt</pre>
<p>Create the reFind configuration file:</p>
<pre>printf 'menuentry "sd" {\n\tvolume "sd"\n\tloader /boot/vmlinuz\n\tinitrd /boot/core.gz\n\toptions "libata.allow_tpm=1 quiet base loglevel=3 waitusb=10 superuser mse-iso rssd-fw-update rssd-fwdir=/opt/firmware rssd-model=MX200 mse-poweroff \n}\n"' &gt; /mnt/EFI/BOOT/refind.conf</pre>
<p>Copy files from ISO</p>

[dc_coder_84]

I would like to see a comparison of application startup times with these SSDs/HDDs. I assume the optane SSD will perform best because of the hight speed at low queue dephts.

Most Nvme SSDs can only perform best at high queue dephs and I wonder if applications like Firefox can use this potential? Wouldn't that be a good field to improve performance for Firefox?

[ThoreauHD]

If samsung, crucial, or intel make one of these, I'll consider it. I need the iso firmware. Running some windows horse shit to update drive firmware ain't gonna happen at home or work.

[Space Heater]

No, that doesn't mean the companies on that list support firmware updates on Linux. They support firmware encapsulation which is an entirely different beast. It's platform agnostic because it's using the board's firmware to update itself.

Yes I am fully aware they are using standard UEFI capsule updates, but that doesn't change the fact that you can start this process on Linux rather than having to launch a Windows-only executable to start the process. You're arguing pure semantics, when it comes to many products the *only* reasonable way to start the capsule update is through a win32 program, so it really doesn't matter that it's using something that is platform-agnostic afterwards.