Importing BlackParrot into RTLMeter #15
Conversation
| LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND | ||
| ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT | ||
| (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS | ||
| SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. No newline at end of file |
There was a problem hiding this comment.
Nit: please add a \n to the end of all the license files so git diffs etc don't complain.
| origin: | ||
| - repository: https://github.com/black-parrot/black-parrot.git | ||
| revision: b57d05165bb44a062aaa5b9675af928f847f2fee | ||
| licenses: |
There was a problem hiding this comment.
Nit: Please strip trailing white space in descriptor.yaml
| - src/bp/bp_nonsynth_if_verif.sv | ||
| - src/bp/bp_nonsynth_nbf_loader.sv | ||
| - src/bp/bp_nonsynth_perf.sv | ||
| - src/bp/bp_monitor.cpp |
There was a problem hiding this comment.
When building with a Verialtor configured with --enale-ccwarn, gcc 13.3.0 reports an "infinite recursion" warning in bp_monitor.cpp, however, looks like this file is not required, so hopefully it can just be removed.
In general please double check if there are any other files that are not actually needed for the design (no need to try everything, just whatever you can eyeball).
| - -pvalue+max_instr_p=10000000 | ||
| - -pvalue+max_cycle_p=10000000 | ||
| - -pvalue+stall_cycles_p=100000 | ||
| - -pvalue+halt_instr_p=1000 | ||
| - -pvalue+heartbeat_instr_p=10000 | ||
| - -pvalue+sim_clock_period_p=10 | ||
| - -pvalue+sim_reset_cycles_lo_p=0 | ||
| - -pvalue+sim_reset_cycles_hi_p=20 | ||
| - -pvalue+tb_clock_period_p=2 | ||
| - -pvalue+tb_reset_cycles_lo_p=0 | ||
| - -pvalue+tb_reset_cycles_hi_p=50 |
There was a problem hiding this comment.
| - -pvalue+max_instr_p=10000000 | |
| - -pvalue+max_cycle_p=10000000 | |
| - -pvalue+stall_cycles_p=100000 | |
| - -pvalue+halt_instr_p=1000 | |
| - -pvalue+heartbeat_instr_p=10000 | |
| - -pvalue+sim_clock_period_p=10 | |
| - -pvalue+sim_reset_cycles_lo_p=0 | |
| - -pvalue+sim_reset_cycles_hi_p=20 | |
| - -pvalue+tb_clock_period_p=2 | |
| - -pvalue+tb_reset_cycles_lo_p=0 | |
| - -pvalue+tb_reset_cycles_hi_p=50 | |
| - -Gmax_instr_p=10000000 | |
| - -Gmax_cycle_p=10000000 | |
| - -Gstall_cycles_p=100000 | |
| - -Ghalt_instr_p=1000 | |
| - -Gheartbeat_instr_p=10000 | |
| - -Gsim_clock_period_p=10 | |
| - -Gsim_reset_cycles_lo_p=0 | |
| - -Gsim_reset_cycles_hi_p=20 | |
| - -Gtb_clock_period_p=2 | |
| - -Gtb_reset_cycles_lo_p=0 | |
| - -Gtb_reset_cycles_hi_p=50 |
| - src/basejump_stl/bsg_mem_dma.hpp | ||
| topModule: testbench | ||
| mainClock: testbench.dut_clk | ||
| verilatorArgs: |
There was a problem hiding this comment.
Please add --autoflush to verilatorArgs so lines are printed when ready. This makes progress easier to see.
|
|
||
|
No newline at end of file |
There was a problem hiding this comment.
Nit: single closing \n in file so git doesn't complain.
| initial | ||
| begin | ||
| if (width_p*els_p > 256) | ||
| $display("## %L: instantiating width_p=%d, els_p=%d, read_write_same_addr_p=%d, harden_p=%d (%m)" | ||
| ,width_p,els_p,read_write_same_addr_p,harden_p); | ||
| end |
There was a problem hiding this comment.
These initial $displays are a bit noisy, please remove them so when running we only see the reset banners, then the processor outputs.
% git grep -l "instantiating" | cat
src/basejump_stl/bsg_launch_sync_sync.sv
src/basejump_stl/bsg_mem_1r1w.sv
src/basejump_stl/bsg_mem_1r1w_one_hot.sv
src/basejump_stl/bsg_mem_1r1w_sync.sv
src/basejump_stl/bsg_mem_1r1w_sync_synth.sv
src/basejump_stl/bsg_mem_1rw_sync.sv
src/basejump_stl/bsg_mem_1rw_sync_mask_write_bit.sv
src/basejump_stl/bsg_mem_1rw_sync_mask_write_byte.sv
src/basejump_stl/bsg_mem_1rw_sync_synth.sv
src/basejump_stl/bsg_mem_2r1w_sync.sv
src/basejump_stl/bsg_mem_2r1w_sync_synth.sv
src/basejump_stl/bsg_mem_3r1w_sync.sv
src/basejump_stl/bsg_mem_3r1w_sync_synth.sv
src/basejump_stl/bsg_sync_sync.sv
src/basejump_stl/bsg_tag_master.sv
There was a problem hiding this comment.
Note: a few small other prints before the reset banner are fine, e.g. "Initializing mem", let's just remove the many "instantiated" ones to reduce noise.
|
Also note the 4x4 config seems to fail on dhrystone, but the post check script accepts it: |
2 participants
This PR serves to import multi-core BlackParrot designs into RTLMeter, with various multi-core workloads.