Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[VTA] Refactor to increase platform coverage (Ultra96 etc.) #3496

Merged
merged 64 commits into from
Jul 29, 2019

Conversation

tmoreau89
Copy link
Contributor

@tmoreau89 tmoreau89 commented Jul 4, 2019

Refactoring the VTA codebase to allow for increased FPGA platform coverage.

Currently WIP.

edit: Summary of code changes

  • Transition to Vivado 2018.3
  • Small refactor of VTA HLS code base: index derivation of instruction fields is now implicit; we use bitfields instead to get an automated layout
  • vta_config.json is now simplified - all FPGA specific parameters are now derived from the target in pkg_config.py, and BLOCK_IN and BLOCK_OUT are fused into BLOCK
  • Parameter derivation is mostly done in pkg_config.py to simplify argument passing to tcl scripts
  • Adding preliminary Ultra96 support, although e2e support is pending on heterogeneous runtime support

@tqchen
Copy link
Member

tqchen commented Jul 23, 2019

@tmoreau89 can you act on this PR as it is a bit outdated

@tmoreau89
Copy link
Contributor Author

@tqchen sure, it's still WIP; as there have been multiple merges in the meantime so I'll do a rebase tomorrow

@tmoreau89
Copy link
Contributor Author

@tqchen just rebased to tvm master

@tmoreau89
Copy link
Contributor Author

@vegaluisjose @liangfu @tqchen @liangfu please review, thanks!

vta/hardware/xilinx/sim/vta_test.cc Show resolved Hide resolved
vta/include/vta/hw_spec.h Outdated Show resolved Hide resolved
vta/python/vta/environment.py Outdated Show resolved Hide resolved
@tmoreau89 tmoreau89 requested review from tqchen and jroesch July 28, 2019 08:21
@jroesch jroesch merged commit f55609b into apache:master Jul 29, 2019
@SrivastavaKshitij
Copy link

SrivastavaKshitij commented Aug 2, 2019

Please refer to the issue mentioned above, Build fails after this pull request was merged to the master. I did a quick test and TVM builds successfully using commit related #3590 that landed on master right before this pull request was merged.

wweic pushed a commit to wweic/tvm that referenced this pull request Aug 9, 2019
)

* hardware refactor for increased FPGA coverage, small optimizations

* fix header

* cleaning up parameters that won't be needed for now

* streamlining makefile, and simplifying tcl scripts

* moving parameter derivation into pkg_config.py, keeping tcl scripts lightweight

* refactoring tcl script to avoid global variables

* deriving AXI signals in pkg_config.py

* unifying address map definition for hardware and software drivers

* single channel design for ultra96 to simplify build

* enable alu by default, no mul opcode for now

* hardware fix

* new bitstream; vta version

* avoid error when env variable is not set

* ultra96 cleanup

* further cleaning up tcl script for bitstream generation

* preliminary rpc server support on ultra96

* rpc server tracker scripts

* ultra96 ldflag

* ultra96 support

* ultra96 support

* cleanup line

* cmake support for ultra96

* simplify memory instantiation

* cleaning up IP parameter initialization

* fix queue instantiation

* 2019.1 transition

* fix macro def

* removing bus width from config

* cleanup

* fix

* turning off testing for now

* cleanup ultra96 ps insantiation

* minor refactor

* adding comments

* upgrading to tophub v0.6

* model used in TVM target now refers to a specific version of VTA for better autoTVM scheduling

* revert change due to bug

* rename driver files to be for zynq-type devices

* streamlining address mapping

* unifying register map offset values between driver and hardware generator

* rely on cma library for cache flush/invalidation

* coherence management

* not make buffer packing depend on data types that can be wider than 64bits

* refactor config derivation to minimize free parameters

* fix environment/pkg config interaction

* adding cfg dump property to pkgconfig:

* fix rpc reconfig

* fix spacing

* cleanup

* fix spacing

* long line fix

* fix spacing and lint

* fix line length

* cmake fix

* environment fix

* renaming after pynq since the driver stack relies on the pynq library - see pynq.io

* update doc

* adding parameterization to  name

* space

* removing reg width

* vta RPC

* update doc on how to edit vta_config.json

* fix path

* fix path
wweic pushed a commit to neo-ai/tvm that referenced this pull request Sep 6, 2019
)

* hardware refactor for increased FPGA coverage, small optimizations

* fix header

* cleaning up parameters that won't be needed for now

* streamlining makefile, and simplifying tcl scripts

* moving parameter derivation into pkg_config.py, keeping tcl scripts lightweight

* refactoring tcl script to avoid global variables

* deriving AXI signals in pkg_config.py

* unifying address map definition for hardware and software drivers

* single channel design for ultra96 to simplify build

* enable alu by default, no mul opcode for now

* hardware fix

* new bitstream; vta version

* avoid error when env variable is not set

* ultra96 cleanup

* further cleaning up tcl script for bitstream generation

* preliminary rpc server support on ultra96

* rpc server tracker scripts

* ultra96 ldflag

* ultra96 support

* ultra96 support

* cleanup line

* cmake support for ultra96

* simplify memory instantiation

* cleaning up IP parameter initialization

* fix queue instantiation

* 2019.1 transition

* fix macro def

* removing bus width from config

* cleanup

* fix

* turning off testing for now

* cleanup ultra96 ps insantiation

* minor refactor

* adding comments

* upgrading to tophub v0.6

* model used in TVM target now refers to a specific version of VTA for better autoTVM scheduling

* revert change due to bug

* rename driver files to be for zynq-type devices

* streamlining address mapping

* unifying register map offset values between driver and hardware generator

* rely on cma library for cache flush/invalidation

* coherence management

* not make buffer packing depend on data types that can be wider than 64bits

* refactor config derivation to minimize free parameters

* fix environment/pkg config interaction

* adding cfg dump property to pkgconfig:

* fix rpc reconfig

* fix spacing

* cleanup

* fix spacing

* long line fix

* fix spacing and lint

* fix line length

* cmake fix

* environment fix

* renaming after pynq since the driver stack relies on the pynq library - see pynq.io

* update doc

* adding parameterization to  name

* space

* removing reg width

* vta RPC

* update doc on how to edit vta_config.json

* fix path

* fix path
tqchen pushed a commit to tqchen/tvm that referenced this pull request Mar 29, 2020
)

* hardware refactor for increased FPGA coverage, small optimizations

* fix header

* cleaning up parameters that won't be needed for now

* streamlining makefile, and simplifying tcl scripts

* moving parameter derivation into pkg_config.py, keeping tcl scripts lightweight

* refactoring tcl script to avoid global variables

* deriving AXI signals in pkg_config.py

* unifying address map definition for hardware and software drivers

* single channel design for ultra96 to simplify build

* enable alu by default, no mul opcode for now

* hardware fix

* new bitstream; vta version

* avoid error when env variable is not set

* ultra96 cleanup

* further cleaning up tcl script for bitstream generation

* preliminary rpc server support on ultra96

* rpc server tracker scripts

* ultra96 ldflag

* ultra96 support

* ultra96 support

* cleanup line

* cmake support for ultra96

* simplify memory instantiation

* cleaning up IP parameter initialization

* fix queue instantiation

* 2019.1 transition

* fix macro def

* removing bus width from config

* cleanup

* fix

* turning off testing for now

* cleanup ultra96 ps insantiation

* minor refactor

* adding comments

* upgrading to tophub v0.6

* model used in TVM target now refers to a specific version of VTA for better autoTVM scheduling

* revert change due to bug

* rename driver files to be for zynq-type devices

* streamlining address mapping

* unifying register map offset values between driver and hardware generator

* rely on cma library for cache flush/invalidation

* coherence management

* not make buffer packing depend on data types that can be wider than 64bits

* refactor config derivation to minimize free parameters

* fix environment/pkg config interaction

* adding cfg dump property to pkgconfig:

* fix rpc reconfig

* fix spacing

* cleanup

* fix spacing

* long line fix

* fix spacing and lint

* fix line length

* cmake fix

* environment fix

* renaming after pynq since the driver stack relies on the pynq library - see pynq.io

* update doc

* adding parameterization to  name

* space

* removing reg width

* vta RPC

* update doc on how to edit vta_config.json

* fix path

* fix path
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants