newgrp design_apd
cd /project/mercury
/bin/bash
export SHELL=/bin/bash
. /project/mercury/bin/chip_env -proj_dir PN85.1.dft_iter0 -prog_desc PN85.1
export DROP_DATA=/project/mercury/source/PN85.1/80_blockdrop/PN85.1.dft
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
Here's the list of 17 blocks that are ready for building :
----------------------------------------
hm_ae
hm_be
hm_capt
hm_clip
hm_conv_n
hm_conv_s
hm_dl
hm_efr
hm_fh
hm_grad
hm_icn
hm_iop
hm_l1dl
hm_l1ul
hm_pss
hm_ul
hm_xlgx
Notes :
-----------------
- Netlist REV_TAG : BROADCOM_REV_TAG_PN85_1_2024_04_30
- Floorplan DEF REV_TAG : PN85.1.fp7.09May2024
- Pinpoint server : http://pinpoint.lvn.broadcom.net:9840/Mercury
- Schedule : Please have these blocks complete AVO by May 27, 2024, or sooner if possible.
- Please send me a list/spreadsheet of the blocks and who they will be assigned to.
Floorplanner Notes (from Matthew Michaels):
-----------------------------------------------------
1) All pins should be placed. If you find that this is not the case, please place them and notify +Matthew Michels (APD)
In addition, all pins should be legal, if not please legalize them using "legalizePin", and notify +Matthew Michels (APD)
2) use the following for your VDDM power gridding (all blocks except hm_pss, pcie_ss_pcie_0, & pss_subsys_nokmercury) :
$CHIPROOT/designers/matthew/vddm_power.tcl
This should essentially get run in lieu of the ag_add_power_grid for the blocks with dual rail memories (almost all of them). NOTE : It has at the top of the script a margin variable that it uses to combine macros that are close enough to not get other power grid inserted in between them. I'm not sure what value to set that to exactly - it might need modification by the block builders. Please notify +Matthew Michels (APD) and +Rob Bassett (APD) if these variables ($mergeX, $mergeY) need to be modified.
3) Please run the prep "promote" target - this will run a Calibre DRC task ("cal_drc_base_layers_gc_filled") which should highlight any problems in the power grid structure quickly.
##############################################
newgrp design_apd
cd /project/mercury
/bin/bash
export SHELL=/bin/bash
. /project/mercury/bin/chip_env -proj_dir PN85.1.dft_iter0 -prog_desc PN85.1
export DROP_DATA=/project/mercury/source/PN85.1/80_blockdrop/PN85.1.dft.PCIE_and_ARM_Memory
export MY_BLOCK=<your block name here - see the list below>
export MY_BLOCK_ROOT=$BLOCKPATH/$MY_BLOCK
/project/mercury/bin/mercury_block_setup
cd $GEV_PROJECT_ROOT/user/$USER/$PROJECT_DIR/impl/$MY_BLOCK
Here's the list of 2 blocks that are ready for building :
----------------------------------------
pcie_ss_pcie_0
pss_subsys_nokmercury
Notes :
-----------------
- Netlist REV_TAG : BROADCOM_REV_TAG_PN85_1_2024_04_30
- Floorplan DEF REV_TAG : PN85.1.fp7.09May2024
- Pinpoint server : http://pinpoint.lvn.broadcom.net:9840/Mercury
- Schedule : Please have these blocks complete AVO by June 12, 2024, or sooner if possible.
- Please send me a list/spreadsheet of the blocks and who they will be assigned to.
Floorplanner Notes (from +Matthew Michels (APD) ):
-----------------------------------------------------
1) All pins should be placed. If you find that this is not the case, please place them and notify +Matthew Michels (APD)
In addition, all pins should be legal, if not please legalize them using "legalizePin", and notify +Matthew Michels (APD)
2) use the following for your VDDM power gridding (all blocks except hm_pss, pcie_ss_pcie_0, & pss_subsys_nokmercury) :
$CHIPROOT/designers/matthew/vddm_power.tcl
This should essentially get run in lieu of the ag_add_power_grid for the blocks with dual rail memories (almost all of them). NOTE : It has at the top of the script a margin variable that it uses to combine macros that are close enough to not get other power grid inserted in between them. I'm not sure what value to set that to exactly - it might need modification by the block builders. Please notify +Matthew Michels (APD) and +Rob Bassett (APD) if these variables ($mergeX, $mergeY) need to be modified.
3) Please run the prep "promote" target - this will run a Calibre DRC task ("cal_drc_base_layers_gc_filled") which should highlight any problems in the power grid structure quickly.
Static Timing Analysis Notes (from +Mark Nevin ) :
-------------------------------------------------------------------
The pcie block had issues because of the bussed clocks rx_wordclk[0] and rx_wordclk[1].
I think I have good constraints for these clocks but there is a non-zero chance that innovus or some other tool doesn't like the syntax I have. Please keep an eye out for errors in the flow.
##################################################
newgrp design_apd
cd /project/mercury
/bin/bash
export SHELL=/bin/bash
. /project/mercury/bin/chip_env -proj_dir PN85.1.dft_iter1 -prog_desc PN85.1
export DROP_DATA=/project/mercury/source/PN85.1/80_blockdrop/PN85.1.dft.2024_05_30_clock_uncr
export MY_BLOCK=<your block name here - see the list below>
export MY_BLOCK_ROOT=$BLOCKPATH/$MY_BLOCK
/project/mercury/bin/mercury_block_setup
Here's the list of 4 blocks that are ready for building :
----------------------------------------
hm_conv_n
hm_conv_s
hm_pss
hm_xlgx
All the other block drop files are the same as from the May 13, 2024 block drop (email thread was titled "[Mercury] PN85.1.dft block drop ready").
You should be able to use your best floorplanned version of DEF from those builds for this build and not need any block-level floorplanning.
Please build these in parallel with your previous PN85.1 builds and target an AVO completion by June 13, 2024.
Please respond back to this email thread with the path to the location where you start each of these builds.
#######################################################################HM_AE#####################
The Mercury customer has given us two (2) new netlists that break up the muxing structure in hm_ae.
Because they are wanting to deliver a PN85.1.5 full chip netlist by mid June, they would like to have feedback on this aforementioned change ASAP.
Currently, DFT is working on the netlist, but in parallel, we are hoping that you can run the "noDFT" version of the netlists through our GIANT flow to see if the logic changes help the block's timing performance and buildability.
Can you start builds with the following three (3) netlists?
-rw-rw-r-- 1 ges design_apd 2030342049 Apr 30 11:14 /project/mercury/source/PN85.1/20_flat/hm_ae.v
-rw-rw-r-- 1 ges design_apd 2117438526 Jun 4 15:17 /project/mercury/source/PN85.1_hm_ae_v1/20_flat_block/hm_ae.v
-rw-rw-r-- 1 ges design_apd 2146889740 Jun 4 15:16 /project/mercury/source/PN85.1_hm_ae_v2/20_flat_block/hm_ae.v
I would recommend that you use the following three environment script command lines to set these up for building (plus all the other usual stuff):
. /project/mercury/bin/chip_env -proj_dir PN85.1.nodft_iter0 -prog_desc PN85.1
export DROP_DATA=/project/mercury/source/PN85.1/20_flat/
. /project/mercury/bin/chip_env -proj_dir PN85.1.nodft_hm_ae_v1_iter0 -prog_desc PN85.1
export DROP_DATA=/project/mercury/source/PN85.1_hm_ae_v1/20_flat_block/
. /project/mercury/bin/chip_env -proj_dir PN85.1.nodft_hm_ae_v2_iter0 -prog_desc PN85.1
export DROP_DATA=/project/mercury/source/PN85.1_hm_ae_v2/20_flat_block/