generated from riscv/docs-spec-template
-
Notifications
You must be signed in to change notification settings - Fork 1
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Signed-off-by: dhower-qc <[email protected]>
- Loading branch information
Showing
1 changed file
with
28 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,28 @@ | ||
= Scalar Efficiency SIG 8/5/2024 | ||
|
||
== Attendees | ||
|
||
* Christian Herber (NXP) | ||
* Derek Hower (Qualcomm) | ||
* David Weaver (Akeana) | ||
* Al Martin (Akeana) | ||
* Mehul Shah (Rivos) | ||
* Ana Pazos (Qualcomm) | ||
* Punit Agrawal (Bytedance) | ||
* Olaf Bernstein (Individual) | ||
* Allen Baum (Esperanto) | ||
* Ved Shanbhogue (Rivos) | ||
* Greg Favor (Ventana) | ||
|
||
== Topics | ||
|
||
* 48-bit instructions. Olaf suggested defining new 32-bit instructions in the 48-bit instruction opcode space. | ||
** Christian pointed out 48-bit space may not be as free as it seems, since it may get used with large immediate instructions | ||
** Derek observed that this would buy at most one extra major opcode int 32-bit space | ||
* Allen suggested prefix instructions to extend immediate of following instruction | ||
** Requires extra CSR register that is saved/restored | ||
** May have implementations costs (adds implicit source and dest to following instruction) | ||
*** ...unless you can gaurantee neither an exception or interrupt can happen between prefix | ||
* Derek pointed out that Qualcomm has released a custom RISC-V microcontroller spec | ||
** Qualcomm will release LLVM implementation for instructions, which SE Sig can use. | ||
* Punit will send out results from load/store pair performance on silicon. |