tvm-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From tqchen via TVM Discuss <nore...@discuss.tvm.ai>
Subject [TVM Discuss] [Development] Google lasted work: MLIR Primer
Date Mon, 08 Apr 2019 17:06:43 GMT


Some of the good directions of MLIR like text format are already present in relay.  And a
natural next step would be the unification of relay and tensor level IR to bring a unified
TVM IR.

Note that MLIR did not make the choice of IR design, so we still make a deliberate choice
on how to design the IR. We could use more thoughts on "pattern-matching style pass" vs transformations
on CFG(which we could move to a different thread). Note that both are pattern matching on
different structures. In my experience, Halide's loop nesting IR benefit from certain high-level
information and quick prototyping.  While CFG+SSA is good for codegen(role of LLVM).





---
[Visit Topic](https://discuss.tvm.ai/t/google-lasted-work-mlir-primer/1721/18) to respond.

You are receiving this because you enabled mailing list mode.

To unsubscribe from these emails, [click here](https://discuss.tvm.ai/email/unsubscribe/dc54e79ab911be8385d670b42a22c94b652c1dc83f3d2646e60b385eea7ea747).

Tianqi Chen, UW, Seattle, WA, 98105, United States
http://tracking.discuss.tvm.ai/tracking/unsubscribe?msgid=Xba-UHEegsYDtpfK8ASSEg2
Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message