From Atul :Dec 10, 2024 at 2:36
SoC PrePN99 Release: from Atul
https://drive.google.com/file/d/14E0JuRy9wuMi-BllL30X7PHIR69QUdfs/view?usp=drive_link
https://drive.google.com/file/d/1t2aHGlQA9BFP6bdVhD8Fs-hyD2NMM8a7/view?usp=drive_link
newgrp design_apd
cd /project/mercury
/bin/bash
. /project/mercury/bin/new_chip_env -proj_dir prePN99.dft_iter_12212024 -prog_desc prePN99
export DROP_DATA=/project/mercury/source/prePN99/80_blockdrop/prePN99.fp12
export MY_BLOCK=<your block name here - see the list below>
export MY_BLOCK_ROOT=$BLOCKPATH/$MY_BLOCK
/project/mercury/bin/mercury_block_setup
cd $GEV_PROJECT_ROOT/user/$USER/$PROJECT_DIR/impl/$MY_BLOCK
cp -rf /project/mercury/source/prePN99/80_blockdrop/prePN99.fp12/${MY_BLOCK}/${MY_BLOCK}.LEQ_info.tcl scripts/con/.
Date: Sat, Dec 21, 2024 at 2:00 PM
Hello Mahesh and Mercury Block builders,
We have made a block drop for prePN99 here and there are notes below. The main goal is to get DEF back to Nokia so if you could try to do a macro placement so that we can send this to them ASAP.
Regards,
Matthew
INFO [00:21:31] Block Drop Directory: /project/mercury/source/prePN99/80_blockdrop/prePN99.fp12
INFO [00:21:31] Blocks Dropped:
INFO [00:21:31] hm_ae
INFO [00:21:31] hm_be
INFO [00:21:31] hm_clipi
INFO [00:21:31] hm_clipo
INFO [00:21:31] hm_conv_n
INFO [00:21:31] hm_conv_s
INFO [00:21:31] hm_dl
INFO [00:21:31] hm_efr_n
INFO [00:21:31] hm_efr_s
INFO [00:21:31] hm_fh_n
INFO [00:21:31] hm_fh_s
INFO [00:21:31] hm_grad
INFO [00:21:31] hm_iop
INFO [00:21:31] hm_l1dl
INFO [00:21:31] hm_l1ul
INFO [00:21:31] hm_mac
INFO [00:21:31] hm_msix
INFO [00:21:31] hm_pss
INFO [00:21:31] hm_ulcap
INFO [00:21:31] hm_xlgx
INFO [00:21:31] pcie_ss_pcie_0
INFO [00:21:31] pss_subsys_nokmercury
Known Issues/Cavaets:
- There will be no cubby for this. The goal is to get some DEF files back to the customer that they can use for physical synthesis of the final PN99 netlists. We should try to do that ASAP
- There are several outstanding requests for pin changes. It will take too long to have the floorplanner be in the middle of each pin movement discussion. To that end, I would suggest that each block builder use agPinPlanPins/PinPlanner to place the pins in groups where they believe they are needed and that can then be incorporated for the final PN99 pin placement. Santhosh will be taking this over and is free to do some other approach.
- Block Area Request changes need customer sign off. We can't grow the blocks in this chip without also impacting die area. If you believe a block must be closed, the customer has to agree to size the block that way.
- We still have some major work to do on the IVD for the right side of the chip. WE will see if we can get some improved top_context files based on this floorplan
- There are a few blocks that are a little off by size:
Block 'hm_fh_n' (1532.16 x 630.344)
Required Area: 1007201.9098
GCC Area : 33303.0298
Actual Area: 965787.8630 {x-delta (+31.92)} {y-delta (+10.868)}
Delta Area: -41414.0467 (-4.11 %)
Block 'hm_be' (1627.92 x 402.116)
Required Area: 670126.5700
Actual Area: 654612.6787 {x-delta (+63.84)} {y-delta (+10.868)}
Delta Area: -15513.8913 (-2.32 %)
Block 'hm_ae' (2234.4 x 576.004)
Required Area: 1307438.4000
Actual Area: 1284941.8982 {x-delta (+63.84)} {y-delta (+10.868)}
Delta Area: -22496.5018 (-1.72 %)
Block 'hm_efr_s' (1532.16 x 2032.316)
Required Area: 3146884.8498
GCC Area : 33303.0298
Actual Area: 3113833.2826
Delta Area: -33051.5672 (-1.05 %)
AE/BE we are trying to be more aggressive on so we left them untouched. FH_N could possibly grow but not without some bigger floorplan changes and getting this drop out seemed more important. We wanted to make you aware though
- There may be other issues in this drop that need to be reviewed but getting something out seemed better than waiting for perfect in this situation.
Tue, Dec 24, 3:57 AM
Hi Matthew,
After talking with Steve here is the SDC for the missing blocks pss_subsys_nokmercury and the pcie_ss_pcie_0 .
/project/mercury/source/prePN99/70_timing.agnel/block_data/sdc_deliver_pcie_pss_subsys_12232024/
Sat, Jan 11, 12:44 PM
/project/mercury/source/prePN99/70_timing.agnel/block_data/sdc_deliver_fixes_wbt
Wed, Jan 15, 2025 at 6:15 AM
/project/mercury/source/prePN99/70_timing.agnel/block_data/sdc_deliver_hm_mac_clock_uncertainty_fix/
[01/17] (changed the clock groups)
/project/mercury/source/prePN99/70_timing.agnel/block_data/sdc_deliver_all_blocks_01162025
Sat, Jan 18, 2025
/project/mercury/source/prePN99/70_timing.agnel/block_data/sdc_deliver_fix_innovus_sdc_hm_ae_hm_conv_s_hm_conv_n/
Sat, Jan 18, 2025 at 5:11 AM
/project/mercury/source/prePN99/80_blockdrop.hier/prePN99.fp11.dft.hier_conv
[01/25] (added a guard in the SDC to not try to to create the generated clocks if they are already created and also to check that the master clock is not a cdc clock)
/project/mercury/source/prePN99/70_timing.agnel/block_data/sdc_deliver_fix_cdc_issues_clocks_01242025
/project/mercury/source/prePN99/70_timing.agnel_hier/sdc_deliver_fix_cdc_hier_01242025/