Data Transfer Confusion?

Have an AMD desktop machine - RYZEN 7 3800X CPU on a ASUS AMD AM4 ROD StrixX-570 mobo:
2 Gigabyte Aorus 1 TB Gen 4 M2 drives

Also have an Intel Desktop - Intel Core i7-9700K on a ASUS ROG Strix Z390 mobo :
Samsung 970 EVO - 1 TB
Samsung 970 PPRO - 500GB

All 4 drives benchmark well within specs using CrystalDiskMark and ATTO
AMD machine runs at 4.1-4.3 GHz - cores not synced
Intel machine runs at 4.9 GHz - all cores sunced
Below are snapshots of file transfers on each machine - files being transferred are 74 GB Acronis backups.

AMD machine on top - I cannot for the life of me, understand how the transfer rate on both machines can be even close considering the difference in read/write speeds of the drives involved.

Any/all insight appreciated

Samuria said:

When you copy bih files it load the files into cache which is fast once cache is full the rate can alter

Thanks for the reply - I don't understand your answer -

NavyLCDR said:

Don't forget about bus speeds that the chipsets are running at. Northbridge bus for NVMe drives and Southbridge bus for SATA, I think?

If I recall correctly - both are set to 100 - does that sound right?
Lots of uncharted territory here for me.

24c said:

Have an AMD desktop machine - RYZEN 7 3800X CPU on a ASUS AMD AM4 ROD StrixX-570 mobo:
2 Gigabyte Aorus 1 TB Gen 4 M2 drives

Also have an Intel Desktop - Intel Core i7-9700K on a ASUS ROG Strix Z390 mobo :
Samsung 970 EVO - 1 TB
Samsung 970 PPRO - 500GB

All 4 drives benchmark well within specs using CrystalDiskMark and ATTO
AMD machine runs at 4.1-4.3 GHz - cores not synced
Intel machine runs at 4.9 GHz - all cores sunced
Below are snapshots of file transfers on each machine - files being transferred are 74 GB Acronis backups.

AMD machine on top - I cannot for the life of me, understand how the transfer rate on both machines can be even close considering the difference in read/write speeds of the drives involved.

Any/all insight appreciated

Be aware transfer speeds depend on a myriad of things including file/file size, hardware, hardware configuration, OS, background processes to name a few.

Transfer speed (like read/write) is not an exact science but an "up to" estimate where everything in the transfer chain has to be perfect. "Perfect" is typically not the case as explained above because of those influences on transfer speeds.

You're also looking at two different builds where overall performance can be equalized. Meaning your tests would be more valid if both drives were tested on the same hardware

sygnus21 said:

Be aware transfer speeds depend on a myriad of things including file/file size, hardware, hardware configuration, OS, background processes to name a few.
Transfer speed (like read/write) is not an exact science but an "up to" estimate where everything in the transfer chain has to be perfect. "Perfect" is typically not the case as explained above because of those influences on transfer speeds.
You're also looking at two different builds where overall performance can be equalized. Meaning your tests would be more valid if both drives were tested on the same hardware

I agree that "speed" is not an exact science by any means.

The whole point of the AMD build was to use the Gen 4 PCIe speed - the drives in the AMD build all benchmark at the 5000/4400 read/write speeds.
I expected the transfer speed between the two Gen 4 M2 drives to be much faster than the Intel machine with Gen 3 PCIe.

The issue is my lack of knowledge on how data is transferred.


OT - in reference to your signature - Kobe and his daughter, who wanted to play for UConn, were honored by the Women's basketball team:

24c said:

OT - in reference to your signature - Kobe and his daughter were honored by the UConn Women's basketball team

@2 Huskies :)

24c said:

I agree that "speed" is not an exact science by any means.


The whole point of the AMD build was to use the Gen 4 PCIe speed - the drives in the AMD build all benchmark at the 5000/4400 read/write speeds.
I expected the transfer speed between the two Gen 4 M2 drive to be much faster than the Intel machine with Gen 3 PCIe.

The issue is my lack of knowledge on how data is transferred.

My "speed" point is you're comparing two different systems so you've got one of four possible outcomes verses 2 on the same system. That is est one drive, test the other drive. You could compare the other machine but it doesn't support NVMe4 so test is invalid.

on the other machine. But in either case you're looking at the drives themselves and they perform in both systems

The issue is you're comparing two systems (AMD v Intel) at two different bus speeds (SATA v NVMe4). Yes, on paper, the NVNe4 should be killing that SATA drive, but things get equalized due to the different systems and hardware configuration. There's also software considerations. So yeah, while interesting, really isn't a fair test.

My two cents.

I think I get it now....I like the Intel machine better -

24c said:

I think I get it now....I like the Intel machine better -

So you're saying you like Intel system with the slower SSD over the the AMD system with the faster SSD.

And the sake of clarity - my thoughts are not about AMD vs. Intel or SATA v NVMe. It's about testing methodologies. And what I take of the OP's last comment is they prefer one build over the other. Not a statement of which chip is better.

My two cents.