`
`removing redundant files.
`
`Change 94950 on 2003/04/10 by askende@askende_r400linuxmarlboro
`
`removing redundant files. The module are now part of the spi block
`
`Change 94873 on 2003/04/10 by askende@askende_r400_linux_marlboro
`
`releasing the following changes:
`1. creation of the new SPI block
`
`2.
`
`3.
`
`top level changes to support
`
`8 SP instances
`
`tracker changes to support a few IO name changes
`
`Change 94830 on 2003/04/09 by mmantor@mmantor_crayola_linux_orl
`
`<SQ/SX/SP out of order thread completion and remove redundant storage in sp for sq/sx
`
`communitcations some sq cfs bug fixed and texture kill mask generation and other misc
`
`things>
`
`Change 94730 on 2003/04/09 by askende@askende_r400linuxmarlboro
`
`timing fix
`
`Change 93993 on 2003/04/04 by askende@askende_r400_linux_marlboro
`
`checked in a timing fix
`
`Change 93160 on 2003/04/01 by askende@askende_r400_linux_marlboro
`
`checking in a timing (tentative)
`
`fix
`
`Change 93010 on 2003/03/31 by askende@askende_r400linuxmarlboro
`
`change related to
`
`vector engines (7 cycles)
`
`matching the latency of the predicate and kill on both scalar and
`
`Change 92655 on 2003/03/28 by askende@askende_r400linuxmarlboro
`
`changes related to :
`
`1. NaN propagation in case of DOT products
`
`2. qualifying the inf detection on the input of the scalar engine for operand
`
`b with ~singleoperand_instruction
`
`Change 92534 on 2003/03/27 by askende@askende_r400linuxmarlboro
`
`AMD1044_0254089
`
`ATI Ex. 2094
`IPR2023-00922
`Page 1 of 24
`
`
`
`added/updated comments...
`
`Change 92515 on 2003/03/27 by askende@askende_r400linuxmarlboro
`
`changes related to :
`
`l.infinity propagation in case of DOT product
`2.ignoring the NaN inputs in SrcC when dealing with two-operand opcodes
`
`Change 92203 on 2003/03/26 by askende@askende_r400_linux_marlboro
`
`reverting the previous change.
`
`Change 92018 on 2003/03/25 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to detecting an overflow and clamping to max_float in the
`scalar engine multiplier
`
`Change 92000 on 2003/03/25 by hartogs@fl_hartogs
`
`Un-did previous change so that I would have to listen to Skende whine about it.
`
`Change 91979 on 2003/03/25 by hartogs@fl_hartogs
`
`Put the "USEBEHAVEMEM"
`
`(dum_mem_p2) model back in.
`
`Change 91689 on 2003/03/22 by askende@askende_r400_linux_marlboro
`
`reverting a change made to the swizzle logic in sp_macc.v rev.56
`
`Change 91601 on 2003/03/21 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to maskgt opcode
`
`Change 91566 on 2003/03/21 by askende@askende_r400linuxmarlboro
`
`fix related to a
`are
`
`typo in the swizzle logic when different channels of the same operand
`
`selected from different sources (GPR vs. Previous result vs. Constant)
`
`Change 90988 on 2003/03/19 by askende@askende_r400linuxmarlboro
`
`fix related to KILL (MASK) scalar instruction
`
`Change 90972 on 2003/03/19 by askende@askende_r400_linux_marlboro
`
`releasing changes to top level related to SoCBIST
`
`Change 90786 on 2003/03/18 by askende@askende_r400linuxmarlboro
`
`AMD1044_0254090
`
`ATI Ex. 2094
`IPR2023-00922
`Page 2 of 24
`
`
`
`fixes related to:
`
`1l.vector MAX4 instruction
`
`2.TRUNC/FLOOR vector instruction
`
`Change 90624 on 2003/03/17 by askende@askende_r400_linux_marlboro
`
`releasing a change related to NaN propagation in case of one-operand scalar
`instructions
`
`Change 90245 on 2003/03/14 by askende@askende_r400_linux_marlboro
`
`various changes
`
`Change 90171 on 2003/03/14 by askende@askende_r400linuxmarlboro
`
`fixes related to pred_pop instruction
`
`Change 90001 on 2003/03/13 by tien@tien_r400devel_marlboro
`
`emask fixes, EMU still incorrect
`
`Change 89943 on 2003/03/13 by askende@askende_r400_linux_marlboro
`
`releasing fix related to param gen inputs into the interpolators
`
`Change 89848 on 2003/03/12 by askende@askende_r400linuxmarlboro
`
`1. added opcode decoding to detect PREDPOP, INV, CLR and RESTORE in sp_macc.v
`2. routed the results of the above opcodes into the outputs of the scalar engine
`
`Change 89778 on 2003/03/12 by askende@askende_r400linuxmarlboro
`
`fix related to
`
`1. vector pixel kill
`
`2. infinity propagation in case of DOT2ADD vector opcode
`
`Change 89491 on 2003/03/11 by vbhatia@vbhatia_r400_linux_marlboro
`
`Example stimulus generation tests
`
`Change 89330 on 2003/03/10 by askende@askender400linuxmarlboro
`
`change related to flushing to zero the mantissa of the result when the exp <= 0
`
`Change 89222 on 2003/03/09 by askende@askende_r400_linux_marlboro
`
`fix related to SETGT instruction in the vector engine
`
`AMD1044_0254091
`
`ATI Ex. 2094
`IPR2023-00922
`Page 3 of 24
`
`
`
`Change 89131 on 2003/03/07 by askende@askende_r400linuxmarlboro
`
`releasing changes related to
`
`1. CUBE opcode .
`
`2. forcing to zero the mantissa for denorm float results.
`
`Change 88814 on 2003/03/06 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to generation of an R400_FP_NAN when dealing with neg values in
`the case of SQRT instruction
`
`Change 87889 on 2003/03/01 by askende@askende_r400_linux_marlboro
`
`fix related to
`
`1. NaN and Inf when executing MAX4 opcode
`
`Change 87845 on 2003/02/28 by tien@tien_r400develmarlboro
`
`Fix for Nans and Infs
`
`Change 87738 on 2003/02/28 by askende@askende_r400linuxmarlboro
`
`released a fix related to Inf values in case of the DOT product
`
`Change 87200 on 2003/02/26 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to detecting an exponent overflow and clamping to MAXFLOAT at
`the output of the interpolator
`
`Change 87093 on 2003/02/26 by tien@tien_r400develmarlboro
`
`Fixed a port issue in tp
`
`Added max clamp in sptpnorm32
`Fixed max/min clamp in sp_tp_norm32
`Fixed sens list issue in sp_tp_norm32
`
`Change 87010 on 2003/02/26 by askende@askende_r400linuxmarlboro
`
`changes related to
`
`1.timing
`
`2.Nan and Inf in the case of vector FLOOR opcode
`
`Change 86930 on 2003/02/25 by tien@tien_r400develmarlboro
`
`Fix for 1D tex maps
`Fix for formatter
`
`Flopped tp_id
`
`AMD1044_0254092
`
`ATI Ex. 2094
`IPR2023-00922
`Page 4 of 24
`
`
`
`Remove 1'bl from clk_en
`
`Change 86765 on 2003/02/25 by askende@askende_r400linuxmarlboro
`
`submitting changes related to
`
`1. fract vector opcode
`2. kill mask vector opcode
`
`3. DOT handling of NaN and Inf values
`
`Change 86625 on 2003/02/24 by tien@tien_r400devel_marlboro
`
`DXN fixes
`
`24/8 float fixes
`16 float fixes
`
`Change 86557 on 2003/02/24 by tien@tien_r400develmarlboro
`
`248 fixes plus fixes for tp4-tc tests
`
`Change 86440 on 2003/02/24 by asutkows@asutkows_r400sun_marlboro
`
`latest sp.syn file.
`
`Change 86327 on 2003/02/23 by askende@askende_r400_linux_marlboro
`
`fixes related to
`
`1. FLOOR instruction in the scalar engine
`2. Handling of NaN values when executing SQRT instruction
`
`Change 86287 on 2003/02/23 by askende@askende_r400linux_marlboro
`
`releasing fixes related to TRUNC and FLOOR opcodes
`
`Change 86180 on 2003/02/21 by askende@askende_r400linuxmarlboro
`
`change related to a timing fix in the MOVA const address path
`
`Change 86025 on 2003/02/21 by jmarsano@R400_FUSEBOXESJMARSANO
`
`PTA files for sp and sx
`
`Change 85869 on 2003/02/21 by rramsey@rramseycrayolalinuxorl
`
`remove wr address pipelining from visrl,2,3 and just use
`
`address from sq
`
`Change 85834 on 2003/02/21 by tien@tien_r400devel_marlboro
`
`AMD1044_0254093
`
`ATI Ex. 2094
`IPR2023-00922
`Page 5 of 24
`
`
`
`Bug fixes
`
`Change 85804 on 2003/02/20 by askende@askende_r400linuxmarlboro
`
`releasing changes related to vector FLOOR instructions
`
`Change 85357 on 2003/02/19 by askende@askende_r400_linuxmarlboro
`
`submitting the correct
`
`.bvrl that represents what is checked in
`
`as sp_scalar_lut.mc
`
`Change 85314 on 2003/02/19 by askende@askender400linuxmarlboro
`
`fixed a bug related to handling of NaN inputs in case of FRACT instrucions
`
`Change 85154 on 2003/02/19 by askende@askende_r400_linux_marlboro
`
`releasing an experimental timing fix related to the adder path in the scalar
`
`engine
`
`Change 85135 on 2003/02/18 by askende@askende_r400_linux_marlboro
`
`fixes related to :
`
`1. handling of -0.0 vs. +0.0 in the case of scalar MIN/MAX instruction
`2. replaced norm({)
`function with a leading one detection + shift left logic
`
`Change 85128 on 2003/02/18 by tien@tien_r400develmarlboro
`
`Refresh of
`
`.bvrls,
`
`just in case
`
`Moved RF exp adjusts to tp_hicolor
`
`Change 84795 on 2003/02/17 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to handling of -INF when an input into RECIPSORT*
`
`Change 84756 on 2003/02/17 by askende@askende_r400linux_marlboro
`
`fixes related to :
`
`1. NaN and Inf in the scalar adder
`
`2. XY screen coordinates into interpolators
`
`Change 84718 on 2003/02/17 by askende@askender400linuxmarlboro
`
`releasing fixes related to :
`
`1. FLOOR and FRACT scalar opcodes
`
`2.
`
`inserted output flops on the vertex parameter bus going from SX to SP
`
`interpolators
`
`AMD1044_0254094
`
`ATI Ex. 2094
`IPR2023-00922
`Page 6 of 24
`
`
`
`Change 84631 on 2003/02/15 by askende@askende_r400_linuxmarlboro
`
`releasing fixes related to sqrt behavior when invalid inputs as well as
`
`fract on infinity input
`
`Change 84499 on 2003/02/14 by askende@askende_r400_linuxmarlboro
`
`releasing a change related to NaN * 0.0 case and kill mask instrunction
`
`Change 84249 on 2003/02/13 by askende@askende_r400linux_marlboro
`
`releasing a fix related to handling of inf when executing FRACT opcode
`
`Change 84246 on 2003/02/13 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to CNDGE when dealing with -0 vs. +0
`
`Change 84063 on 2003/02/12 by tien@tien_r400devel_marlboro
`
`Fixed shifts for as_16_ 1616 16 modes, both RF and INT
`
`Change 83970 on 2003/02/12 by tien@tien_r400devel_marlboro
`
`16Float to 32Float y and w channel fix
`
`Change 83632 on 2003/02/11 by askende@askende_r400linuxmarlboro
`
`submitting a timing fix related to FRACT opcode path.
`
`Change 83624 on 2003/02/11 by tien@tien_r400develmarlboro
`
`Fixed degamma
`
`Added ports for multisample and hi latency
`
`Change 83448 on 2003/02/11 by tien@tien_r400devel_marlboro
`
`Timing fix for SP
`
`Change 83422 on 2003/02/10 by askende@askende_r400linuxmarlboro
`
`submittig a timing fix
`
`Change 83286 on 2003/02/10 by mmang@mmang_crayola_linux_orl
`
`Delayed vector engine operation predicate result bits by 2 clocks to line up with
`
`scalar engine.
`
`AMD1044_0254095
`
`ATI Ex. 2094
`IPR2023-00922
`Page 7 of 24
`
`
`
`Change 83262 on 2003/02/10 by vbhatia@vbhatia_r400linuxmarlboro
`
`Example input stimulus files
`
`Change 83260 on 2003/02/10 by vbhatia@vbhatia_r400linuxmarlboro
`
`Checkin of a simple testbench for Tien to use in debug which just reads a simple
`
`stimulus input file and outputs the sp_tp_formatter output as defined in the plusargs
`arguments of the simv build. These outputs
`have to be compared with the emulator
`output.
`
`Change 83230 on 2003/02/10 by askende@askender400linuxmarlboro
`
`submitted changes related to the endian ordering of the pred/kill mask bits
`
`Change 83097 on 2003/02/08 by askende@askende_r400_linux_marlboro
`
`fix related to NaN values used as input in FRACT, FLOOR and TRUNC instructions
`
`Change 83091 on 2003/02/08 by askende@askende_r400linuxmarlboro
`
`submitting a change related to MOVA on the scalar engine side
`
`Change 83073 on 2003/02/08 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to float-to-fix conversion on the MOVA address data
`
`Change 82980 on 2003/02/07 by askende@askende_r400_linux_marlboro
`
`releasing precision fixes for SP DOT opcodes
`
`Change 82964 on 2003/02/07 by tien@tien_r400devel_marlboro
`
`Fixes for 16-bit float exp=0 case
`Fixes for exp adjusts for some of the newer formats
`
`Change 82498 on 2003/02/06 by askende@askende_r400_linux_marlboro
`
`precision changes related to MULADD logic
`
`Change 82189 on 2003/02/05 by hartogs@fl_hartogs
`
`Added missing port to instantiation of sp_vector modules to prevent warnings during sim
`load.
`
`Change 82044 on 2003/02/05 by askende@askende_r400_linux_marlboro
`
`submitting a tentative fix related to pred_override usage in spmacc.v
`
`AMD1044_0254096
`
`ATI Ex. 2094
`IPR2023-00922
`Page 8 of 24
`
`
`
`Change 81792 on 2003/02/04 by askende@askende_r400_linuxmarlboro
`
`fixed a bug related to FLOOR of values that have an exponent > 24.
`
`Change 81455 on 2003/02/03 by askende@askende_r400_linux_marlboro
`
`checking in a fix related to RECIP function when dealing with max float values as
`
`inputs.
`
`Change 81348 on 2003/02/02 by askende@askende_r400_linux_marlboro
`
`releasing a timing fix in the sp_macc32_multiply unit.
`
`Change 81253 on 2003/02/01 by askende@askende_r400_linuxmarlboro
`
`submitting a fix related to scalar mul_prev2 opcode
`
`Change 81230 on 2003/01/31 by askende@askende_r400linuxmarlboro
`
`submitting sp changes
`
`Change 81108 on 2003/01/31 by askende@askende_r400_linux_marlboro
`
`backing up changes
`
`Change 80845 on 2003/01/30 by askende@askende_r400linuxmarlboro
`
`releasing a fix related to pred_set_execute latency in the scalar engine
`
`Change 80805 on 2003/01/30 by tien@tien_r400develmarlboro
`
`Updated to sptp stuff
`
`Change 80784 on 2003/01/30 by askende@askende_r400linuxmarlboro
`
`fixed a width mismatch problem
`
`Change 80360 on 2003/01/29 by askende@askende_r400_linux_marlboro
`
`releasing changes related to
`
`1.
`
`interpolation modes
`
`2. signaling_nan and pos_nan handling in the vector engine
`
`Change 80283 on 2003/01/29 by askende@askende_r400_linuxmarlboro
`
`checking in a fix related to the sign of the result when dealing with TRUNC instruction
`
`AMD1044_0254097
`
`ATI Ex. 2094
`IPR2023-00922
`Page 9 of 24
`
`
`
`Change 80281 on 2003/01/29 by askende@askende_r400linux_marlboro
`
`releasing DOT precision fixes
`
`Change 79573 on 2003/01/26 by askende@askende_r400linuxmarlboro
`
`releasing changes to the interpolator logic to fix a timing path
`
`Change 79522 on 2003/01/26 by askende@askende_r400_linux_marlboro
`
`fixed problem with rsq_ieee opcode when input is -0/+0, —/+denorm float.
`
`Change 79519 on 2003/01/26 by askende@askende_r400_linux_marlboro
`
`changes/fixes related to :
`1. cylindrical wrap logic in SX
`
`2. behavior of logieee when dealing with corner values inputs like -0,+0 and
`
`inf
`
`Change 79357 on 2003/01/24 by askende@askende_r400linuxmarlboro
`
`releasing changes to
`
`1. mace multiply exponent overflow detection and clamping to MAX_FLOAT
`2. clamping in the scalar unit
`
`Change 79177 on 2003/01/24 by scamlin@scamlin_crayola_win
`
`virage star test validation
`
`Change 79012 on 2003/01/23 by askende@askende_r400_linux_marlboro
`
`fixed a bug related to flushing the subnorms to zero on the output of the interpolators
`
`Change 78893 on 2003/01/23 by desiree@desireer400sun_marlboro
`
`added WEMA to memory
`
`Change 78839 on 2003/01/23 by desiree@desireer400_sun_marlboro
`
`Ripped out fusebox, related signals, and matched Test Team spec
`
`Change 78826 on 2003/01/23 by desiree@desireer400sun_marlboro
`
`removed duplicate fuseboxes
`
`Change 78754 on 2003/01/22 by subad@subad_r400_ linux_marlboro
`
`fixed DXTC non-degammaed and made sptpformatter 8B for DXTC cases
`
`AMD1044_0254098
`
`ATI Ex. 2094
`IPR2023-00922
`Page 10 of 24
`
`
`
`Change 78691 on 2003/01/22 by desiree@desireer400sun_marlboro
`
`deleting from this level. Mems are under /vector and /interpolator
`
`Change 78686 on 2003/01/22 by desiree@desiree_r400sun_marlboro
`
`updated for virage/3220
`
`Change 78685 on 2003/01/22 by desiree@desiree_r400_sun_marlboro
`
`updated for virage/3220
`
`Change 78683 on 2003/01/22 by desiree@desireer400sun_marlboro
`
`updated for virage/3220
`
`Change 78681 on 2003/01/22 by desiree@desireer400sun_marlboro
`
`updated for virage/3220
`
`Change 78679 on
`
`2003/01/22 by desiree@desiree_r400sun_marlboro
`
`updated name of
`
`fusebox for virage/3220
`
`Change 78677 on
`
`2003/01/22 by desiree@desiree_r400sun_marlboro
`
`changed name of
`
`fusebox for virage/3220
`
`Change 78646 on
`
`2003/01/22 by desiree@desireer400sun_marlboro
`
`changed fusebox
`
`name to sp_rffusebox
`
`Change 78617 on
`
`2003/01/22 by askende@askende_r400linuxmarlboro
`
`fix related to the clamp logic in both of the modules below
`
`Change 78461 on 2003/01/22 by askende@askende_r400_linuxmarlboro
`
`fixes related to:
`
`1. mask/kill logic
`
`2. alloc/dealloc logic in Sx
`
`Change 78336 on 2003/01/21 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to sp_const_sel
`select logic.
`
`in the alu constant
`
`AMD1044_0254099
`
`ATI Ex. 2094
`IPR2023-00922
`Page 11 of 24
`
`
`
`Change 78247 on 2003/01/21 by hartogs@fl_hartogs
`
`Added this port to the instance “uscalar" to file sp/vector/spvector.v.
`.OPREDSET_EXECUTE(),
`// TODO -- added unconnected output port to avoid
`warnings during sim load
`
`Apparently the spscalarlut module added an output port (single bit).
`
`Also added these ports to the instance "u_sq_tex_ctl_flow_seq" in sq.v
`-ais_update(),//
`listed unconnected port to avoid load warning
`-ais_thread_id({),//
`listed unconnected port to avoid load warning
`ais_update is single bit, and ais_thread_id is 6 bits.
`
`Change 77934 on 2003/01/20 by askende@askende_r400_linux_marlboro
`
`releasing changes related to
`1. square root precision
`
`2. scalar add exp underflow detection
`
`3. constant waterfalling logic
`
`Change 77800 on 2003/01/19 by askende@askende_r400linuxmarlboro
`
`submitting a new sp.v that includes the new sq_sp_const_sel bus.
`
`Change 77796 on 2003/01/19 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to SET opcodes
`
`Change 77784 on 2003/01/19 by askende@askende_r400_linux_marlboro
`
`changes related to :
`
`l. predicate logic
`
`2. set opcodes in the vector unit
`
`Change 77734 on 2003/01/18 by askende@askende_r400linuxmarlboro
`
`changes related to sqrt precision
`
`Change 77654 on 2003/01/17 by askende@askende_r400linuxmarlboro
`
`fixed a problem related to tp_sp_gpr_cmask and tpsp_gprdst buses being
`pipelined when
`not needed
`
`Change 77529 on 2003/01/17 by askende@askende_r400_linux_marlboro
`
`submitting changes related to DOT4 precision and some internal port renaming/clean up
`
`Change 77383 on 2003/01/17 by subad@subad_r400linuxmarlboro
`
`AMD1044_0254100
`
`ATI Ex. 2094
`IPR2023-00922
`Page 12 of 24
`
`
`
`fixed syntax error
`
`Change 77365 on 2003/01/17 by subad@subad_r400_linuxmarlboro
`
`fix for 8B alpha channel for DXTC
`
`Change 77211 on 2003/01/16 by desiree@desireer400_sun_marlboro
`
`SP control file that conforms to virage/3220 and project conventions
`
`Change 77163 on 2003/01/16 by askende@askender400linuxmarlboro
`
`changes related to :
`
`top level SQ-SP predicate/kill vs. mova interface.
`1.
`2. fix related to propagation of the sign in the case of MUL or MULADD opcodes
`
`when argument a or b are inf inputs.
`
`Change 76994 on 2003/01/15 by desiree@desireer400sun_marlboro
`
`new processor from Jan. 14th virage star compiler release
`
`Change 76992 on 2003/01/15 by fhsien@fhsien_r400_linux_marlboro
`
`Change MESALL parameter for more virage memory
`
`Change 76989 on 2003/01/15 by desiree@desireer400sun_marlboro
`
`fusebox verilog from new Jan. 14th compiler version
`
`Change 76987 on 2003/01/15 by desiree@desireer400sun_marlboro
`
`fusebox generated with new Jan. 14 release of virage compiler
`
`Change 76895 on 2003/01/15 by askende@askende_r400linuxmarlboro
`
`MOVA instruction changes related to performance optimization
`
`Change 76871 on 2003/01/15 by askende@askende_r400_linux_marlboro
`
`releasing MOVA optimization changes
`
`Change 76558 on 2003/01/14 by askende@askende_r400_linux_marlboro
`
`submitting changes related to
`
`1. illegal inputs generating a r400fpnan for opcode like LOG, SQRT,
`RECIP_SORT etc.
`2. decreasing the multiply result precision in the vector muladd
`
`AMD1044_0254101
`
`ATI Ex. 2094
`IPR2023-00922
`Page 13 of 24
`
`
`
`Change 76112 on 2003/01/13 by askende@askende_r400_linuxmarlboro
`
`releasing changes related to
`
`1.scalar engine
`
`2. export quad index logic in SX
`
`Change 75637 on 2003/01/09 by hartogs@fl_hartogs
`
`Added missing port signals with empty parentheses to avoid warnings during simulation
`load.
`
`Change 75599 on 2003/01/09 by askende@askende_r400_linux_marlboro
`
`releasing a fix related to scalar engine add opcode precision
`
`Change 75549 on 2003/01/09 by askende@askende_r400linuxmarlboro
`
`releasing changes related to interpolator precision
`
`Change 75503 on 2003/01/09 by askende@askende_r400_linux_marlboro
`
`releasing changes related to
`
`1. scalar add precision
`2. sx unit id assignemet
`
`Change 75343 on 2003/01/08 by dougd@dougd_r400linuxmarlboro
`
`changes for vgt-sq-sp vertex vector loading performance improvement
`
`Change 75317 on 2003/01/08 by askende@askende_r400_linux_marlboro
`
`releasing the following changes:
`
`1. fix related to scalar_add precision
`2. fix related to mova vector opcode and SP_SQconst_addr interface
`
`Change 74443 on 2003/01/03 by askende@askende_r400linuxmarlboro
`
`fix related to FRACT opcode in the vector engine
`
`Change 74373 on 2003/01/03 by askende@askender400linuxmarlboro
`
`releasing the HW precision changes for MACC32 of the vector unit
`
`Change 74346 on 2003/01/03 by tien@tien_r400develmarlboro
`
`Fixed mantissa shift for exp=0..4
`
`AMD1044_0254102
`
`ATI Ex. 2094
`IPR2023-00922
`Page 14 of 24
`
`
`
`Change 73944 on 2002/12/31 by askende@askende_r400_linuxmarlboro
`
`submitting fixes related NaN * zero = zero
`
`Change 73924 on 2002/12/31 by tien@tien_r400devel_marlboro
`
`Redid 655 fix
`
`Change 73499 on 2002/12/27 by tien@tien_r400devel_marlboro
`
`Fixed 655 z channel shift
`
`Change 73428 on 2002/12/26 by askende@askende_r400linuxmarlboro
`
`submitting a fix related to MUL opcode when one of the arguments is INF
`
`Change 73264 on 2002/12/23 by askende@askende_r400linuxmarlboro
`
`fix related to max4 instruction.
`
`Change 73218 on 2002/12/23 by askende@askende_r400linuxmarlboro
`
`removed a redundant signal declaration
`
`Change 73185 on 2002/12/23 by tien@tien_r400devel_marlboro
`
`Misc fixes, makes some tests run better, but not pass completely
`
`Change 73170 on 2002/12/23 by askende@askende_r400_linux_marlboro
`
`connected the "writeaddr" bus to the star memory module
`
`Change 73029 on 2002/12/21 by askende@askende_r400linuxmarlboro
`
`splitted the macc32 unit into
`
`1.
`
`sp_macce32.me(bvr1)
`
`2. sp_macc32_multiply.mc(bvr1l)
`
`Change 72833 on 2002/12/20 by askende@askende_r400linuxmarlboro
`
`fix a bug related to inf and/or nan inputs in the SP (sp_macc32.mc) module
`
`Change 72408 on 2002/12/19 by askende@askende_r400_linux_marlboro
`
`releasing changes related to handling of Inf inputs
`
`Change 72079 on 2002/12/18 by askende@askende_r400linuxmarlboro
`
`AMD1044_0254103
`
`ATI Ex. 2094
`IPR2023-00922
`Page 15 of 24
`
`
`
`submitting ij performance optimization changes
`
`Change 71476 on 2002/12/16 by askende@askende_r400linuxmarlboro
`
`submitting changes related to
`1.
`fog logic
`
`2.
`
`trunc/floor logic
`
`Change 71361 on 2002/12/16 by desiree@desireer400sun_marlboro
`
`fixed TST_awt_enable and STAR_rfrecen from linter output
`
`Change 71314 on 2002/12/15 by askende@askende_r400linuxmarlboro
`
`l.releasing changes related to chip.tree
`
`2.fixed the log2 function behavior when srce_a = NaN
`
`Change 70793 on 2002/12/12 by askende@askende_r400linuxmarlboro
`
`releasing changes related to :
`
`1. corner case value for muladd (min_plus_1, min_plus2)
`2. constant waterfalling and mova instructions
`
`Change 70218 on 2002/12/10 by desiree@desireer400sun_marlboro
`
`adding fuse box ctme file
`
`Change 70217 on 2002/12/10 by desiree@desireer400sun_marlboro
`
`fuse box related code
`
`Change 70216 on 2002/12/10 by desiree@desireer400sun_marlboro
`
`processor for star memories added for the first time
`
`Change 70210 on 2002/12/10 by desiree@desiree_r400sun_marlboro
`
`changes related to fuse-box, processor, delay-chain and other test
`
`signals related to star memories
`
`Change 70205 on 2002/12/10 by desiree@desireer400sun_marlboro
`
`changes related to processor,
`star memories
`
`fuse-box, delay chain and other test signals related to
`
`Change 70201 on 2002/12/10 by desiree@desireer400sun_marlboro
`
`AMD1044_0254104
`
`ATI Ex. 2094
`IPR2023-00922
`Page 16 of 24
`
`
`
`changes related to processor,
`star memories
`
`fuse-box, delay chain and other test signals related to
`
`Change 70200 on 2002/12/10 by
`
`desiree@desireer400sun_marlboro
`
`changes related to processor,
`star memories
`
`fuse-box, delay chain and other test signals related to
`
`Change 70199 on 2002/12/10 by
`
`desiree@desiree_r400sun_marlboro
`
`changes related to processor,
`
`delay chain,
`
`fuse box and other test
`
`signals related to star memories
`
`Change 70198 on 2002/12/10 by desiree@desireer400sun_marlboro
`
`includes processor,fuse-box, delay-chain and other test signals related
`to star memories
`
`Change 70193 on 2002/12/10 by
`
`askende@askende_r400linux_marlboro
`
`submitting top level changes
`
`Change 69512 on 2002/12/08 by
`
`askende@askende_r400_linux_marlboro
`
`fixed corner cases for nan and inf on lookup-table like opcodes
`
`Change 69332 on 2002/12/06 by
`
`askende@askende_r400linuxmarlboro
`
`submitting changes related to
`
`TRUNC opcode
`
`Change 69231 on 2002/12/06 by
`
`askende@askende_r400linux_marlboro
`
`checking in changes
`
`Change 68613 on 2002/12/04 by
`
`askende@askende_r400_linuxmarlboro
`
`submitting changes....
`
`1. Have implemeted the
`2. Fixed a corner case
`
`new behavior of FRACT
`
`related MulAdd when dealing with very small input values
`
`Change 68561 on 2002/12/04 by
`
`tien@tien_r400devel_marlboro
`
`Changed size of tpc_fifos fetch RAM
`Changed ports form TP/TC to SP to get formats to work properly
`
`Change 68147 on 2002/12/03 by
`
`askende@askende_r400linux_marlboro
`
`submitting changes related to
`
`TRUNC and FRACT implementation
`
`AMD1044_0254105
`
`ATI Ex. 2094
`IPR2023-00922
`Page 17 of 24
`
`
`
`Change 67739 on 2002/12/01 by askende@askende_r400_linuxmarlboro
`
`checking in changes related to *_FF and * IEEE version of some of the scalar
`opcodes
`
`Change 67716 on 2002/12/01 by askende@askende_r400_linuxmarlboro
`
`releasing changes related to NaN and Inf inputs in the scalar engine
`
`Change 67677 on 2002/11/30 by askende@askende_r400linux_marlboro
`
`releasing changes
`
`Change 67620 on 2002/11/29 by askende@askende_r400_linuxmarlboro
`
`releasing changes related to support special FP values like NaN and Inf
`
`Change 67379 on 2002/11/27 by askende@askende_r400linuxmarlboro
`
`submitting changes related to LOG opcodes behavior on special case inputs (NaNs and
`Infs)
`
`Change 66775 on 2002/11/25 by askende@askende_r400_linux_marlboro
`
`submitting changes related to special value inputs (NaNs and Inf)
`as well as mace unit
`
`in the interpolators
`
`Change 66531 on 2002/11/24 by askende@askende_r400_linux_marlboro
`
`checking in a change related to TRUNC opcode
`
`Change 66513 on 2002/11/24 by askende@askende_r400linuxmarlboro
`
`releasing changes related to NaN and Inf special input values
`
`Change 66366 on 2002/11/22 by askende@askende_r400_linux_marlboro
`
`checking in changes related to expected math results when dealing with NaNs
`
`Change 66183 on 2002/11/22 by askende@askende_r400_linux_marlboro
`
`submitting a fix related to DOT2ADD opcode...the only DOT2ADD that we currently have is
`
`now passing
`
`Change 65681 on 2002/11/20 by askende@askende_r400_linux_marlboro
`
`submitting a fix related to coissu-ed exports
`
`AMD1044_0254106
`
`ATI Ex. 2094
`IPR2023-00922
`Page 18 of 24
`
`
`
`Change 65669 on 2002/11/20 by askende@askende_r400_linuxmarlboro
`
`submitting fixes related to :
`
`1. co-issued export instructions
`
`2.
`
`trunc instruction operating on very large numbers
`
`Change 65046 on 2002/11/18 by askende@askende_r400_linux_marlboro
`
`added scan IO at the top level
`
`Change 64965 on 2002/11/18 by askende@askender400linuxmarlboro
`
`submitting an interpolation sign detection fix when overflowing out of the final add
`
`Change 64865 on 2002/11/18 by pmitchel@pmitchelentiredepot_win
`
`mv files for Desiree
`
`Change 64819 on 2002/11/18 by desiree@desireer400sun_marlboro
`
`fixed star memory - comma was missing
`
`Change 64717 on 2002/11/17 by askende@askende_r400_linux_marlboro
`
`submitting a fix for trunc and fract opcode
`
`Change 64565 on 2002/11/15 by askende@askende_r400_linux_marlboro
`
`submitting changes related to star memories
`
`Change 64553 on 2002/11/15 by desiree@desireer400sun_marlboro
`
`added wire declaration for vss
`
`Change 64538 on 2002/11/15 by askende@askende_r400linux_marlboro
`
`submitting star memory related changes
`
`Change 64526 on 2002/11/15 by askende@askende_r400linuxmarlboro
`
`submitting changes related to star memories
`
`Change 64507 on 2002/11/15 by desiree@desireer400sun_marlboro
`
`submitting star memory related changes
`
`Change 64502 on 2002/11/15 by desiree@desireer400sunmarlboro
`
`AMD1044_0254107
`
`ATI Ex. 2094
`IPR2023-00922
`Page 19 of 24
`
`
`
`submitting star memory related changes
`
`Change 64498 on 2002/11/15 by desiree@desireer400sun_marlboro
`
`submitting star memory related changes
`
`Change 64496 on 2002/11/15 by desiree@desireer400_sun_marlboro
`
`submiting star memory related logic
`
`Change 64495 on 2002/11/15 by desiree@desireer400sun_marlboro
`
`releasing the star memeory related logic
`
`Change 64474 on 2002/11/15 by askende@askende_r400_linux_marlboro
`
`backing up changes
`
`Change 64067 on 2002/11/14 by askende@askende_r400linuxmarlboro
`
`submitting the dot3 fixes in the sp_macc32 module
`
`Change 63979 on 2002/11/14 by tien@tien_r400devel_marlboro
`
`Area reduction and big fixes for 32-bit normalizers
`
`Change 63951 on 2002/11/14 by askende@askende_r400_linux_marlboro
`
`removed a duplicated assignement of signal osptpdata
`
`Change 63849 on 2002/11/13 by askende@askende_r400linuxmarlboro
`
`l.area reduction changes
`2.fixing DOT products
`
`Change 63185 on 2002/11/11 by askende@askende_r400_linux_marlboro
`
`fixed a bug introduced during the performance optimizations
`
`Change 62245 on 2002/11/07 by askende@askende_r400_linux_marlboro
`
`submitting changes related to pvps and int_gpr_wr_en logic driven by
`new SQ changes
`
`Change 61921 on 2002/11/06 by vromaker@vromakerr400_linux_marlboro
`
`changed "pvps" to “pspv" in the write enable signal names going to the 4 maccs
`
`AMD1044_0254108
`
`ATI Ex. 2094
`IPR2023-00922
`Page 20 of 24
`
`
`
`Change 61767 on 2002/11/05 by askende@askende_r400_linuxmarlboro
`
`fixed a problem related to gpr_int_wr_ena
`
`Change 61276 on 2002/11/03 by askende@askende_r400_linux_marlboro
`
`checking in changes related to scalar FRACT instruction
`
`Change 61259 on 2002/11/03
`
`by
`
`askende@askende_r400_linuxmarlboro
`
`backing up changes
`
`Change 61225 on 2002/11/02
`
`by
`
`askende@askende_r400linuxmarlboro
`
`backing up changes
`
`Change 61222 on 2002/11/02
`
`by
`
`askende@askende_r400linuxmarlboro
`
`backing up changes
`
`Change 60881 on 2002/10/31
`
`by
`
`askende@askende_r400linuxmarlboro
`
`fixed a scalar input modifier
`executed
`
`related bug in the case when a SUB instruction is being
`
`Change 60794 on 2002/10/31
`
`by
`
`askende@askende_r400linuxmarlboro
`
`submitting macc32 fixes
`
`Change 60576 on 2002/10/30
`
`by
`
`askende@askende_r400linux_marlboro
`
`increased the precision in
`
`the adder by adding two gard bits
`
`Change 60185 on 2002/10/29
`
`by askende@askende_r400_linuxmarlboro
`
`fixed a bug related to the
`
`order of precedence of the input modifiers
`
`on the arguments
`
`Change 59960 on 2002/10/29
`
`submitting changes related
`the mace unit
`
`by
`
`to
`
`askende@askender400linuxmarlboro
`
`underflow conditions in the exponent of
`
`Change 59854 on 2002/10/28
`
`by
`
`askende@askende_r400linuxmarlboro
`
`fixed a problem related to
`
`scalar SETGT instruction
`
`AMD1044_0254109
`
`ATI Ex. 2094
`IPR2023-00922
`Page 21 of 24
`
`
`
`Change 59623 on 2002/10/27 by askende@askende_r400linux_marlboro
`
`changes related to :
`1. substract corner case
`
`2. SUBPREV scalar opcode control argument selection overwrite
`
`Change 59587 on 2002/10/27 by askende@askende_r400_linuxmarlboro
`
`fixed the argument selection of the red channel
`
`in the case of scalar instructions
`
`that use the PS as input on red channel
`
`Change 59372 on 2002/10/25 by askende@askender400linuxmarlboro
`
`fixes related to
`
`1. PS detection and argument selection logic
`
`2. SQ_SP_stall usage
`
`Change 59214 on 2002/10/24 by askende@askende_r400linuxmarlboro
`
`fixed a problem related to expdiff when denormalizing to the largest
`exponent.
`
`Change 58952 on 2002/10/23 by askende@askende_r400_linux_marlboro
`
`changes related to signed magnitude math vs.
`
`two's complement math
`
`Change 58780 on 2002/10/23 by askende@askende_r400linuxmarlboro
`
`added code to check for exponent overflow condition
`
`resulting into max_float_ieee value
`
`Change 58756 on 2002/10/22 by askende@askende_r400_linuxmarlboro
`
`fixed a bug related to input modifier ABS
`
`Change 58724 on 2002/10/22 by askende@askende_r400linux_marlboro
`
`fixed scalar implementation of TRUNC opcode
`
`Change 58683 on 2002/10/22 by brianféma_bfavela
`
`Updated for new files
`
`Change 58209 on 2002/10/20 by askende@askende_r400_linux_marlboro
`
`backing up SET* opcode changes
`
`Change 57806 on 2002/10/17 by askende@askende_r400linuxmarlboro
`
`AMD1044_0254110
`
`ATI Ex. 2094
`IPR2023-00922
`Page 22 of 24
`
`
`
`1. fixed a bug related to SUB scalar opcode
`2. added support for the SET* scalar opcodes
`
`Change 57570 on 2002/10/16 by askende@askende_r400linuxmarlboro
`
`submitting fixes related to scalar engine opcodes: MIN, MAX
`
`Change 57203 on 2002/10/15 by askende@askende_r400_linux_marlboro
`
`fixed a latency problem with the scalar engine.
`
`at this point r400sppixscalaradd_* test do pass.
`
`Change 56816 on 2002/10/13 by askende@askende_r400linuxmarlb