MAG Tomahawk X870 M2_1 and M2_2 not achieving advertised speed

leipp124c02a1

New member
Joined
May 11, 2025
Messages
24
tl;dr;
Devices in both M2_1 and M2_2 ports show connecting at 5.0x4.
Each device only reaches 4.7GB/s which is not even gen 4.0x4 max speed of 7GB/s.
I'd expect closer to 14GB/s.

details:
BIOS: latest A44
linux kernel: tried various 6.12.n and also the latest 6.14.6
2x Crucial T705 gen5 NVMe in M2_1 and M2_2 ports.
lspci -vv shows both connected at gen 5.0x4, 32GT/s link speeds.
Testing with dd if=/dev/nvme0n1 of=/dev/null bs=4096 count=10M shows a mere 4.7GB/s on either device.
Testing with hdparm -tT [--direct] /dev/nvme0n1 shows similar slow speeds.
I checked the temps before, during, and after each test run and they never exceeded the 40C range (I think 44C was the highest temp I saw).
They should see close to 14GB/s.
Meanwhile on my older board my older gen 4.0x4 NVMe sees 7GB/s as expected, connected at 4.0x4 16GT/s link speed.
I've tried disabling M2_2, it had no effect on M2_1.
Setting either or both to gen4 speeds they show correctly as 16GT/s link speed and still show only 4.7GB/s performance.
I tried disabling the PCI ASPM, no change.
Not sure what else to try.
 
Here correction the performance of Crucial T705 on M2_1 slot as below
While that's much better it's still too slow. It should be nearly 14000 MB/s. 25% too slow is a lot of missing performance.
 

Attachments

  • Screenshot (1).png
    Screenshot (1).png
    50.7 KB · Views: 47
RMA'd the board. Got it "repaired" and returned.
Reinstalled everything.
Ran all the exact same tests again.
Got the exact same results again.

Samsung 980 Pro 2TB in M2_4 slot: 1.9GB/s. Linux shows it connected at 16GT/s (4.0x4).
Crucial T705 4TB in M2_1 and M2_2 slots: 4.6GB/s. Linux shows them both connecting at 32GT/s (5.0x4)

Meanwhile my old machine can see 6.5GB/s on the Samsung 980 Pro 2TB connected at 16GT/s (4.0x4)....

Something is very wrong still.
 
While that's much better it's still too slow. It should be nearly 14000 MB/s. 25% too slow is a lot of missing performance.
Just to be very clear here. Unless you're doing massive video editing with mega-gigabyte-sized video files and only just gaming and doing the odd thing here or there, you will never, and I mean ever see anywhere near those speeds. So in regards to just gaming, you are not losing any performance at all. It takes a long time for a file to ramp up to the fastest speeds, and that's a huge multi-gigabyte file. Those speeds are not instant. Games load much smaller files, and cannot accomplish the speeds due to the amount of stopping and starting on small game files.

Because of this, I only use SATA SSD's, since the performance with M.2 vs. SATA in gaming, is very little to sometimes no gain. It blows my mind that there are so many posts complaining about not getting "advertised speeds", and not realizing none of you will ever see anywhere near those speeds as far as gaming is concerned. Build your PC and enjoy the performance gains from the new hardware, and not try to wring the last few percent of performance out of something that you'll never notice or see.
 
There's a huge difference between SATA and NVMe speeds even in gaming. Even the old NVMes are 10x faster than SATA, The newest are nearly 30x faster. If you can't notice that difference I dunno what to tell you.

Every single build I've done over the last 20 years I've been able to sanity-check the IO performance before loading the OS. HDD, SSD, and NVMe have always been able to demonstrate performance very near their theoretical maximum with a simple direct streaming read test. It's a great way to just verify that everything is working as it should.

This is the first motherboard I've ever encountered where the newest, latest+greatest technology couldn't even produce performance equal to the prior generation parts. This is not just some "oh you'll never notice thing". I definitely can notice. I don't play games and I regularly perform activities where the IO speed is very, /very/ noticeable, and I regularly see speeds close to the advertised maximums. There's a reason I chose this motherboard and it's not living up to its advertised capabilities.
 
Last edited:
There's a huge difference between SATA and NVMe speeds even in gaming. Even the old NVMes are 10x faster than SATA, The newest are nearly 30x faster. If you can't notice that difference I dunno what to tell you.
I'm going to post a link below the proves you completely wrong, but I am very glad you said that. It tells me you do not understand how NVME drives work. To dumb it down as much as I can: You have to have sustained read/write of a single large file to get anywhere near "advertised" speeds. The speeds start out slow, and ramp up to "advertised" speeds on that one file. And it can only do this with a constant, non-stop read/write of a single massive file because it's not starting and stopping over and over like game files. When you read/write tons of very small files, like loading game files, the speed never gets anywhere near "advertised" because the system doesn't have enough time to get to those speeds having to continuously start and stop for each file.

Ergooooo, as far as gaming is concerned, you will never, ever, ever see anywhere near "advertised" speeds because it is impossible with loading of small game files. You may see some faster speeds in games that have large gigabyte+ files or maps. And loading games is sometimes a bit faster on NVME than SSD But beyond that, no matter what you think or think you perceive, you won't be seeing those advertised speeds in gaming. For instance, I can load a Cyberpunk 2077 save in literally about 5 seconds from my SATA SSD. So unless you can load a save instantly with zero wait time, then your claimed "10x" faster is completely false.

And since you haven't seen any benchmarks, I think my post and these videos sum it up pretty well:

 
Last edited:
Back
Top