AMD RDNA 2 graphics card clues found in ROCm firmware update

by Mark Tyson on 23 September 2020, 11:11

Tags: AMD (NYSE:AMD)

Quick Link: HEXUS.net/qaeoif

Add to My Vault: x

The Radeon open compute (ROCm) firmware was recently updated and has revealed some interesting information nuggets. Redditor stblr spotted the changes to the firmware file which according to their analysis reveals previously unknown info about an AMD GPU codenamed Navy Flounder.

So, what exactly is revealed by this ROCm v3.8 Linux firmware file? It contains the reference and values table as reproduced below:

Redditor stblr explains that interested parties can derive the number of Compute Units (CUs) in the particular GPU by multiplying the top three values together. Another key value indicated by these tables is the memory bus width indicated by the gc_num_tccs value (but there are sometimes exceptions to this value relationship).

Weighing up the above we get the following data:

Sienna Cichlid (Big Navi or Navi 21)

Navy Flounder (Navi 22 or 23)

80 compute units

40 compute units

256-bit GDDR6 memory

192-bit GDDR6 memory

 

If each CU has 64 stream processors, as was the case with RDNA 1 GPUs, then Navy Flounder will be pretty similar in basic configuration to the Radeon RX 5700 XT but with the new architecture and of course it will be positioned at a new competitive level against the latest announced and expected Nvidia Ampere GPUs.

For next-gen console reference, the Xbox Series X SoC features 52 RDNA 2 CUs at 1.83GHz, meanwhile the PlayStation 5 SoC features 36 RDNA 2 CUs at 2.23GHz.

Source: Reddit via Tom's Hardware.



HEXUS Forums :: 18 Comments

Login with Forum Account

Don't have an account? Register today!
Rumours of much less bandwidth are true then… but 80 cu's
Gddr6 or gddr6x ??
Good news,,, well sort off, at my age a month pass by like ( snaps fingers )
And worst of all, this too will probably trigger a really bad case of shopaholic even if i am not going to buy a GFX anytime soon.

Or that is, if AMD do a trade in, then i might have to say BB to my Red devil 5700 XT and somehow sell the unused alphacool cooler for it too. :-)
lumireleon
Gddr6 or gddr6x ??

GDDR6 non X - NV grabbed all of that after AMD bagged TSMC 7nm

Also it seems that the GDDR6X is one of the main reasons for higher power useage and heat
3dcandy
Rumours of much less bandwidth are true then… but 80 cu's

It depends if they've managed to do something like what Nvidia has done their new memory controllers by being able to do 4 signals per clock instead of just 2.

Else there is a big issue with memory starvation as Vega was super memory hungry, that's another reason why it really needed HBM.

To have a huge big “game changer” card then starve its memory will be nuts…