• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Samsung 9100 Pro 2 TB

W1zzard

Administrator
Staff member
Joined
May 14, 2004
Messages
28,814 (3.74/day)
Processor Ryzen 7 5700X
Memory 48 GB
Video Card(s) RTX 4080
Storage 2x HDD RAID 1, 3x M.2 NVMe
Display(s) 30" 2560x1600 + 19" 1280x1024
Software Windows 10 64-bit
The Samsung 9100 Pro is the company's new PCI-Express 5.0 flagship. It reaches transfer speeds of over 14 GB/s and our review confirms that this is the fastest Gen 5 SSD you can buy right now. Energy efficiency is also improved, which helps with heat output, but you still need a dedicated heatsink to avoid throttling completely.

Show full review
 
the after SLC cache writing is disappointing but how did they manage to lose against old SATA drives with a "PRO" Gen 5 Flagship?!
compression-winrar.png
 
Thanks for the review. This will be my next NVMe.
 
Thanks for the review W1zzard!
Nice speeds for a nice price!
1743000271135.png


Why do you list that as a con while there is a with heatsink version is available just like with the 980 PRO and 990 PRO?
The without heatsink is for those who have premium mainboards or AIC adaptor.

Looks like a nice drive for OS :D
 
Last edited:
the after SLC cache writing is disappointing but how did they manage to lose against old SATA drives with a "PRO" Gen 5 Flagship?!
View attachment 391692
This makes no sense… Can someone explain these results? The SLC cache here is actually bigger than the previous Samsung 900 series SSDs, so how is it that in this test this driver is significantly slower, while also winning most of the other benchmarks?

Seems almost like it needs to be retested to me
 
Last edited:
This makes no sense. Can someone explain these results?
Samsung have always sucked at decompressing archives, and this drive is no exception.

It's only one test, and one that's often heavily CPU bottlenecked, so not terribly relevant.
 
Samsung have always sucked at decompressing archives, and this drive is no exception.

It's only one test, and one that's often heavily CPU bottlenecked, so not terribly relevant.
‘Always sucked’ is quite immature not scientific at all, and we’re dealing with data here. It needs to have a reason and meaning.
It could be a bad test. For example, how is it that the rest of the 900 series faired better in this specific test with less cache and lost in most other tests?
 
Thanks to support for the blazing fast PCIe Gen 5 interface, you can reach transfer speeds of up to 14.8 GB/s.

**goes on to test transfer speeds that barely break the 10GB/s barrier**

Edit: Oh wait, it says "you can reach", it's not about the drive. My bad.
 
Thanks for the review W1zzard!
Nice speeds for a nice price!
View attachment 391694

Why do you list that as a con while there is a with heatsink version is available just like with the 980 PRO and 990 PRO?
The without heatsink is for those who have premium mainboards or AIC adaptor.

Looks like a nice drive for OS :D
Some of us in the workstation market view no heatsink as a Pro too. I've bought heatsink drives that were (on sale for cheaper than the alternative o_O) only to then have to shuck them to install in PCIe add-in cards.
 
A nice drive, but too expensive for a basic gamer or PC user. Though people who need this performance, don't need to check the price tag.
 
PCI-E 3.0 SSD's are gold for the PC gaming. Still PCI-E 5.0 SSD's are useless and expensive...
Even SATA SSDs are fine for 99% of the games.
 
For example, how is it that the rest of the 900 series faired better in this specific test with less cache and lost in most other tests?
I have no idea. Each test was run 7 times with reboots in-between: 40.18, 43.47, 39.51, 38.54, 38.72, 39.15, 38.80

**goes on to test transfer speeds that barely break the 10GB/s barrier**

Edit: Oh wait, it says "you can reach", it's not about the drive. My bad.
Check the CDM screenshot
 
Check the CDM screenshot
Does that mean you need to reach QD8 to get 14GB/s? Because no other synthetic test comes close.
 
I have no idea. Each test was run 7 times with reboots in-between: 40.18, 43.47, 39.51, 38.54, 38.72, 39.15, 38.80


Check the CDM screenshot
Thanks for answering, and I always like your other reviews - but as an outlet that many people look up to for accurate reviews (and as people of science!) we should understand the numbers and the results rather than showcasing them in case there’s an error.
As you’re incredibly experienced and still confused yourself on these results - seems like something fishy is going on.
So, I’m not saying you didn’t do those 7 tests - but maybe contact Samsung to see what they had to say? Maybe there was an issue with your specific batch? Maybe it’s worth checking the SSD in a different test bench?
 
Does that mean you need to reach QD8 to get 14GB/s? Because no other synthetic test comes close.
Correct, but this is the case for all SSDs, all vendors report speeds at higher queue depth

So, I’m not saying you didn’t do those 7 tests - but maybe contact Samsung to see what they had to say? Maybe there was an issue with your specific batch? Maybe it’s worth checking the SSD in a different test bench?
Let me erase it, and reinstall the testing OS image and do some compression runs. Maybe it helps freeing up some extra space
 
Correct, but this is the case for all SSDs, all vendors report speeds at higher queue depth
Oh, I know that. Just trying to understand. I thought queues would share the bandwidth, I didn't realize one queue cannot saturate the bus.
 
"World's fastest SSD" of the drives TPU has tested. Both the Crucial T705 and Micron 4600 are arguably better performance wise and don't suffer as severe performance drop when the write cache runs out.

The Crucial T705 of which is significantly cheaper.
 
‘Always sucked’ is quite immature not scientific at all, and we’re dealing with data here. It needs to have a reason and meaning.
It could be a bad test. For example, how is it that the rest of the 900 series faired better in this specific test with less cache and lost in most other tests?
Immature?

Samsung SSD controllers have always been far behind other popular vendors in decompression benchmarks, going back at least a decade. If you want reason and meaning, go and interview some Samsung engineers. Feel free to speculate, but it won't be anything more than speculation which is equally unscientific and presumably unacceptable to you?

I'm just saying what I see across 15+ years of benchmarks from here and places like Anandtech, TechReport, Tom's - and if you don't like casual language then perhaps you ought to stay away from internet forums.
 
Sata became too bad, you can get good ssd with ddr cache, but it costs like pci-e 4.0 nvme.
True, my point was mostly that if you have your old SATA drives still in use, they're fine.
 
Back
Top