EDIT DEC-2-2024: If anyone else is having this issue, try removing the Solidigm driver (from all Solidigm devices) through "Device Manager", reboot & let the Windows default driver install, if you have no isssues under the Windows driver, then just use that driver.
For me, this process resulted in more than 10x improvement in random write speeds, and an end to the pausing phenomenon.
IF like me you experienced odd happenings after loading the Windows default driver: unable to attach files to email, email taking a LONG time to open, web pages taking minutes to load, etc.... then reinstall the Solidigm driver & reboot. For me, I had one or two minor things linger until I did an extra reboot.
However, the re-loaded Solidigm drivers do not represent much of an improvement over the 18 year old Windows NVMe driver.
CDM results for Windows 11 driver VS CDM results for Solidigm driver reload - best results in bold:
[Read] SEQ 1MiB (Q= 8, T= 1):7064.165 MB/s [ 6736.9 IOPS] < 1185.91 us> VS 7012.478 MB/s [ 6687.6 IOPS] < 1195.18 us>
RND 4KiB (Q= 32, T=16): 2088.919 MB/s [ 509990.0 IOPS] < 971.46 us> VS 2236.633 MB/s [ 546053.0 IOPS] < 936.75 us>
[Write] SEQ 1MiB (Q= 8, T= 1):2251.351 MB/s [ 2147.1 IOPS] < 3714.17 us> VS 2629.038 MB/s [ 2507.2 IOPS] < 3186.71 us>
RND 4KiB (Q= 32, T=16): 1419.572 MB/s [ 346575.2 IOPS] < 1455.20 us> VS 1315.498 MB/s [ 321166.5 IOPS] < 1530.87 us>
========================end edit======================
As can be seen by the CDM benchmark, my drive is performing very poorly. Any suggestions? I'm using the Solidigm drivers and I just went through and made sure that all drivers are current. If anything, I think updating drivers made it worse....
Here are the drive details:
System specs: Gigabyte Z690 Gaming X DDR4 motherboard * Intel i9-13900KS CPU with Kraken X63 AIO * WIN 11 PRO 64bit * MSI 4070 Ti Gaming X Trio GPU * Corsair Vengeance 64GB(4x16GB) RAM * Solidigm P44 Pro 2TB M.2 NVMe SSD (Win11 Pro) + Intel 665p 1TB M.2 NVMe (storage) + Samsung 990 Pro 2TB M.2 NVMe (storage) + Samsung 980 1TB M.2 NVMe (Ubuntu) * Seasonic Vertex GX-1200 1200W ATX 3.0 PSU * FANS: Bitfenix Spectre Pro 200mm front intake + 2x Noctua NF-A14 Industrial PPC 140mm exhaust on Kraken X63 radiator
Here's the empty Solidigm P44 Pro drive performance for comparison - very significant write performance difference between an empty drive and a less than 20% full drive:
Dear chaosdsm,
Thank you for reaching out to Solidigm. We are sorry to hear you are experiencing issues with your drive. Could you kindly perform the following steps:
Drive Optimization:
- Confirm the drive is using the optimal connection;
- Drives under test should be connected to the system as a secondary (non OS) drive;
- Run secure erase (Important Note: Always back up data regularly and before proceeding as this process removes all data from the drive.)
- Download & install Solidigm™ Storage Tool and select “Secure Erase” if it’s a NVMe Solidigm™ client SSD;
- Run “Performance Booster”
- After having followed the above-mentioned steps, perform the benchmark process using CrystalDiskMark with the following steps:
- Choose correct SSD type under “Setting” menu
- Choose “Peak Performance” in “Profile” menu
- Select the partition/drive needs to be tested
- Save the results in a text file. This will show the MB/s and the corresponding IOPS results
If you continue to experience any issues with your drive, please contact us via the following link to create a case and we will reach out to you shortly. https://support.solidigm.com/en-US/anonymous-createcase/
Kind regards.
Gleb
Solidigm Customer Support
Drive Optimization:
1. Confirm the drive is using the optimal connection; > it is in the primary nvme slot designated by the motherboard manufacturer.
2. Drives under test should be connected to the system as a secondary (non OS) drive; > No... but I can boot into Ubuntu to do drive testing...
3. Run secure erase > NO, NO, NO, NO - this is my primary drive running Windows.
Download & install Solidigm™ Storage Tool and select “Secure Erase” if it’s a NVMe Solidigm™ client SSD; > I guess you're not too familiar with your own software.... the 2nd photo I posted is clipped from Solidigm Storage Tool 2.0.300..... I'll be sure to include full screenshots from now on....
4. Run “Performance Booster” > Doesn't work, only my Intel 665p drive shows up under Performance Booster - It has been this way in every version of Solidigm Storage Tool.
Here's my list of attached storage devices:
And the List of drives available for Performance Booster
After having followed the above-mentioned steps, perform the benchmark process using CrystalDiskMark with the following steps:
Choose correct SSD type under “Setting” menu
Choose “Peak Performance” in “Profile” menu
Select the partition/drive needs to be tested
Save the results in a text file. This will show the MB/s and the corresponding IOPS results
Peak performance is almost as useless as benchmarking an empty drive. Real-World Performance is what matters....
Here's the peak performance though....
And here's a "Real-World usage" (simulated) test... really shows horrible performance, i.e. modern rotational hard drive like write performance on the IOPS and access time, heck, even the random write The only thing truly remarkable (meaning good) with the "real-world" version is the access time is orders of magnitude faster.
Dear chaosdsm,
Thank you for your detailed feedback and for undertaking the recommended steps to optimize your Solidigm SSD's performance. We apologize that the issues persist despite these efforts.
To assist you more effectively, we would like to invite you to create a support ticket. Please follow this link to create your support case: https://support.solidigm.com/en-US/anonymous-createcase/
Kind regards.
Gleb
Solidigm Customer Support
