Merge branch 'master' of github.com:RRZE-HPC/osaca

osaca/data/bdw.yml

@@ -170,18 +170,6 @@ instruction_forms:
   port_pressure: [[3, '0156']]
   throughput: 0.75
   uops: 3
-- name: '{store} SHRD'
-  operands:
-  - class: immediate
-    imd: int
-  - class: register
-    name: gpr
-  - class: register
-    name: gpr
-  latency: 3
-  port_pressure: [[1, '1']]
-  throughput: 1.0
-  uops: 1
 - name: wait
   operands: []
   latency: 1
@@ -222,6 +210,62 @@ instruction_forms:
   port_pressure: [[1, '06']]
   throughput: 0.5
   uops: 1
+- name: cbtw  # AT&T version of CBW
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cwtl  # AT&T version of CWDE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cwtd  # AT&T version of CWD
+  operands: []
+  latency: 2
+  port_pressure: [[1, '0156'], [1, '06']]
+  throughput: 0.75
+  uops: 2
+- name: cltd  # AT&T version of CDQ
+  operands: []
+  latency: 1
+  port_pressure: [[1, '06']]
+  throughput: 0.5
+  uops: 1
+- name: cltq  # AT&T version of CDQE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cqtq  # AT&T version of CQO
+  operands: []
+  latency: 1
+  port_pressure: [[1, '06']]
+  throughput: 0.5
+  uops: 1
+- name: [cmpeqsd, cmpltsd, cmplesd, cmpunordsd, cmpneqsd, cmpnltsd, cmpnlesd, cmpordsd]
+  # same as CMPSD xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
+- name: [cmpeqss, cmpltss, cmpless, cmpunordss, cmpneqss, cmpnltss, cmpnless, cmpordss]
+  # same as CMPSS xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
 - name: cmpps
   operands:
   - class: immediate
@@ -681,18 +725,38 @@ instruction_forms:
   port_pressure: [[1, '23'], [1, [2D, 3D]]]
   throughput: 0.5
   uops: 2
-
-
-- name: '{store} MOVDQA'
+- name: [vcmpeqsd, vcmpltsd, vcmplesd, vcmpunordsd, vcmpneqsd, vcmpnltsd, vcmpnlesd, vcmpordsd,
+         vcmpeq_uqsd, vcmpngesd, vcmpngtsd, vcmpfalsesd, vcmpneq_oqsd, vcmpgesd, vcmpgtsd,
+         vcmptruesd, vcmpeq_ossd, vcmplt_oqsd, vcmple_oqsd, vcmpunord_ssd, vcmpneq_ussd,
+         vcmpnlt_uqsd, vcmpnle_uqsd, vcmpord_ssd, vcmpeq_ussd, vcmpnge_uqsd, vcmpngt_uqsd,
+         vcmpfalse_ossd, vcmpneq_ossd, vcmpge_oqsd, vcmpgt_oqsd, vcmptrue_ussd]
+  # same as VCMPSD xmm, xmm, xmm, imd
   operands:
   - class: register
     name: xmm
   - class: register
     name: xmm
-  latency: 0
-  port_pressure: [[1, '015']]
-  throughput: 0.3333333333333333
-  uops: 1
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
+- name: [vcmpeqss, vcmpltss, vcmpless, vcmpunordss, vcmpneqss, vcmpnltss, vcmpnless, vcmpordss,
+         vcmpeq_uqss, vcmpngess, vcmpngtss, vcmpfalsess, vcmpneq_oqss, vcmpgess, vcmpgtss,
+         vcmptruess, vcmpeq_osss, vcmplt_oqss, vcmple_oqss, vcmpunord_sss, vcmpneq_usss,
+         vcmpnlt_uqss, vcmpnle_uqss, vcmpord_sss, vcmpeq_usss, vcmpnge_uqss, vcmpngt_uqss,
+         vcmpfalse_osss, vcmpneq_osss, vcmpge_oqss, vcmpgt_oqss, vcmptrue_usss]
+  # same as VCMPSS xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
 - name: vcvtss2sd
   operands:
   - class: register
@@ -28500,6 +28564,17 @@ instruction_forms:
   throughput: 0.25                                          # ./generate_mov_entries.py bdw
   uops: 1                                                   # ./generate_mov_entries.py bdw
                                                             # ./generate_mov_entries.py bdw
+- name: movabs                                              # ./generate_mov_entries.py bdw
+  operands:                                                 # ./generate_mov_entries.py bdw
+  - class: immediate                                        # ./generate_mov_entries.py bdw
+    imd: int                                                # ./generate_mov_entries.py bdw
+  - class: register                                         # ./generate_mov_entries.py bdw
+    name: gpr                                               # ./generate_mov_entries.py bdw
+  latency: 1                                                # ./generate_mov_entries.py bdw
+  port_pressure: [[1, '0156']]                              # ./generate_mov_entries.py bdw
+  throughput: 0.25                                          # ./generate_mov_entries.py bdw
+  uops: 1                                                   # ./generate_mov_entries.py bdw
+                                                            # ./generate_mov_entries.py bdw
 - name: mov  # with store, simple AGU                       # ./generate_mov_entries.py bdw
   operands:                                                 # ./generate_mov_entries.py bdw
   - class: immediate                                        # ./generate_mov_entries.py bdw

osaca/data/csx.yml

@@ -209,6 +209,42 @@ instruction_forms:
   port_pressure: [[1, '01']]  # uops.info (measured)
   throughput: 0.5
   uops: 1
+- name: cbtw  # AT&T version of CBW
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cwtl  # AT&T version of CWDE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cwtd  # AT&T version of CWD
+  operands: []
+  latency: 2
+  port_pressure: [[1, '0156'], [1, '06']]
+  throughput: 0.75
+  uops: 2
+- name: cltd  # AT&T version of CDQ
+  operands: []
+  latency: 1
+  port_pressure: [[1, '06']]
+  throughput: 0.5
+  uops: 1
+- name: cltq  # AT&T version of CDQE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cqtq  # AT&T version of CQO
+  operands: []
+  latency: 1
+  port_pressure: [[1, '06']]
+  throughput: 0.5
+  uops: 1
 - name: cmp
   operands:
   - class: register
@@ -228,6 +264,26 @@ instruction_forms:
   port_pressure: [[1, '0156']]
   throughput: 0.25
   uops: 1
+- name: [cmpeqsd, cmpltsd, cmplesd, cmpunordsd, cmpneqsd, cmpnltsd, cmpnlesd, cmpordsd]
+  # same as CMPSD xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '015']]
+  throughput: 0.3333333333333333
+- name: [cmpeqss, cmpltss, cmpless, cmpunordss, cmpneqss, cmpnltss, cmpnless, cmpordss]
+  # same as CMPSS xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 4
+  port_pressure: [[1, '015']]
+  throughput: 0.3333333333333333
 - name: cmppd
   operands:
   - class: immediate
@@ -526,6 +582,38 @@ instruction_forms:
   throughput: 0.5
   latency: 4.0  # 	1"*"p01
   port_pressure: [[1, '01']]
+- name: [vcmpeqsd, vcmpltsd, vcmplesd, vcmpunordsd, vcmpneqsd, vcmpnltsd, vcmpnlesd, vcmpordsd,
+         vcmpeq_uqsd, vcmpngesd, vcmpngtsd, vcmpfalsesd, vcmpneq_oqsd, vcmpgesd, vcmpgtsd,
+         vcmptruesd, vcmpeq_ossd, vcmplt_oqsd, vcmple_oqsd, vcmpunord_ssd, vcmpneq_ussd,
+         vcmpnlt_uqsd, vcmpnle_uqsd, vcmpord_ssd, vcmpeq_ussd, vcmpnge_uqsd, vcmpngt_uqsd,
+         vcmpfalse_ossd, vcmpneq_ossd, vcmpge_oqsd, vcmpgt_oqsd, vcmptrue_ussd]
+  # same as VCMPSD xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 4
+  port_pressure: [[1, '01']]
+  throughput: 1
+- name: [vcmpeqss, vcmpltss, vcmpless, vcmpunordss, vcmpneqss, vcmpnltss, vcmpnless, vcmpordss,
+         vcmpeq_uqss, vcmpngess, vcmpngtss, vcmpfalsess, vcmpneq_oqss, vcmpgess, vcmpgtss,
+         vcmptruess, vcmpeq_osss, vcmplt_oqss, vcmple_oqss, vcmpunord_sss, vcmpneq_usss,
+         vcmpnlt_uqss, vcmpnle_uqss, vcmpord_sss, vcmpeq_usss, vcmpnge_uqss, vcmpngt_uqss,
+         vcmpfalse_osss, vcmpneq_osss, vcmpge_oqss, vcmpgt_oqss, vcmptrue_usss]
+  # same as VCMPSS xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 4
+  port_pressure: [[1, '01']]
+  throughput: 1
 - name: vcvtsi2ss
   operands:
   - class: register
@@ -627,39 +715,6 @@ instruction_forms:
   throughput: 0.5
   latency: 4.0  # 	1"*"p01
   port_pressure: [[1, '05']]
-- name: vpsrld  # JH: assumed from SKX
-  operands:
-  - class: register
-    name: xmm
-  - class: register
-    name: xmm
-  - class: register
-    name: xmm
-  latency: 1
-  port_pressure: [[1, '0'], [1, '15']]
-  throughput: 1
-- name: VXORPS  # JH: assumed from SKX
-  operands:
-  - class: register
-    name: ymm
-  - class: register
-    name: ymm
-  - class: register
-    name: ymm
-  latency: 1
-  port_pressure: [[1, '015']]
-  throughput: 0.3333333333333333
-- name: VXORPD  # JH: assumed from SKX
-  operands:
-  - class: register
-    name: ymm
-  - class: register
-    name: ymm
-  - class: register
-    name: ymm
-  latency: 1
-  port_pressure: [[1, '015']]
-  throughput: 0.3333333333333333
 - name: wait
   operands: []
   latency: 1
@@ -82603,6 +82658,17 @@ instruction_forms:
   throughput: 0.25                                          # ./generate_mov_entries.py csx
   uops: 1                                                   # ./generate_mov_entries.py csx
                                                             # ./generate_mov_entries.py csx
+- name: movabs                                              # ./generate_mov_entries.py csx
+  operands:                                                 # ./generate_mov_entries.py csx
+  - class: immediate                                        # ./generate_mov_entries.py csx
+    imd: int                                                # ./generate_mov_entries.py csx
+  - class: register                                         # ./generate_mov_entries.py csx
+    name: gpr                                               # ./generate_mov_entries.py csx
+  latency: 1                                                # ./generate_mov_entries.py csx
+  port_pressure: [[1, '0156']]                              # ./generate_mov_entries.py csx
+  throughput: 0.25                                          # ./generate_mov_entries.py csx
+  uops: 1                                                   # ./generate_mov_entries.py csx
+                                                            # ./generate_mov_entries.py csx
 - name: mov  # with store, simple AGU                       # ./generate_mov_entries.py csx
   operands:                                                 # ./generate_mov_entries.py csx
   - class: immediate                                        # ./generate_mov_entries.py csx

osaca/data/generate_mov_entries.py

@@ -148,6 +148,7 @@ snb_mov_instructions = [
     ('mov mem gpr', ('', 0)),
     ('mov imd gpr', ('1*p015', 1)),
     ('mov imd mem', ('', 0)),
+    ('movabs imd gpr', ('1*p015', 1)),  # AT&T version
 
     # https://www.felixcloutier.com/x86/movapd
     ('movapd xmm xmm', ('1*p5', 1)),
@@ -533,6 +534,7 @@ hsw_mov_instructions = list(OrderedDict(ivb_mov_instructions + [
     # https://www.felixcloutier.com/x86/mov
     ('mov imd gpr', ('1*p0156', 1)),
     ('mov gpr gpr', ('1*p0156', 1)),
+    ('movabs imd gpr', ('1*p0156', 1)),  # AT&T version
 
     # https://www.felixcloutier.com/x86/movbe
     ('movbe gpr mem', ('1*p15', 6)),

osaca/data/hsw.yml

@@ -191,6 +191,62 @@ instruction_forms:
   port_pressure: [[1, '0156'], [1, '23']]
   throughput: 0.5
   uops: 3
+- name: cbtw  # AT&T version of CBW
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cwtl  # AT&T version of CWDE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cwtd  # AT&T version of CWD
+  operands: []
+  latency: 2
+  port_pressure: [[1, '0156'], [1, '06']]
+  throughput: 0.75
+  uops: 2
+- name: cltd  # AT&T version of CDQ
+  operands: []
+  latency: 1
+  port_pressure: [[1, '06']]
+  throughput: 0.5
+  uops: 1
+- name: cltq  # AT&T version of CDQE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '0156']]
+  throughput: 0.25
+  uops: 1
+- name: cqtq  # AT&T version of CQO
+  operands: []
+  latency: 1
+  port_pressure: [[1, '06']]
+  throughput: 0.5
+  uops: 1
+- name: [cmpeqsd, cmpltsd, cmplesd, cmpunordsd, cmpneqsd, cmpnltsd, cmpnlesd, cmpordsd]
+  # same as CMPSD xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
+- name: [cmpeqss, cmpltss, cmpless, cmpunordss, cmpneqss, cmpnltss, cmpnless, cmpordss]
+  # same as CMPSS xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
 - name: cmpps
   operands:
   - class: immediate
@@ -375,6 +431,38 @@ instruction_forms:
   port_pressure: [[1, '1']]  # JH: according to uops.info
   throughput: 1.0  # JH: measured on phinally
   uops: 1
+- name: [vcmpeqsd, vcmpltsd, vcmplesd, vcmpunordsd, vcmpneqsd, vcmpnltsd, vcmpnlesd, vcmpordsd,
+         vcmpeq_uqsd, vcmpngesd, vcmpngtsd, vcmpfalsesd, vcmpneq_oqsd, vcmpgesd, vcmpgtsd,
+         vcmptruesd, vcmpeq_ossd, vcmplt_oqsd, vcmple_oqsd, vcmpunord_ssd, vcmpneq_ussd,
+         vcmpnlt_uqsd, vcmpnle_uqsd, vcmpord_ssd, vcmpeq_ussd, vcmpnge_uqsd, vcmpngt_uqsd,
+         vcmpfalse_ossd, vcmpneq_ossd, vcmpge_oqsd, vcmpgt_oqsd, vcmptrue_ussd]
+  # same as VCMPSD xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
+- name: [vcmpeqss, vcmpltss, vcmpless, vcmpunordss, vcmpneqss, vcmpnltss, vcmpnless, vcmpordss,
+         vcmpeq_uqss, vcmpngess, vcmpngtss, vcmpfalsess, vcmpneq_oqss, vcmpgess, vcmpgtss,
+         vcmptruess, vcmpeq_osss, vcmplt_oqss, vcmple_oqss, vcmpunord_sss, vcmpneq_usss,
+         vcmpnlt_uqss, vcmpnle_uqss, vcmpord_sss, vcmpeq_usss, vcmpnge_uqss, vcmpngt_uqss,
+         vcmpfalse_osss, vcmpneq_osss, vcmpge_oqss, vcmpgt_oqss, vcmptrue_usss]
+  # same as VCMPSS xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
 - name: vcvtss2sd
   operands:
   - class: register
@@ -28288,6 +28376,17 @@ instruction_forms:
   throughput: 0.25                                                       # ./generate_mov_entries.py hsw
   uops: 1                                                                # ./generate_mov_entries.py hsw
                                                                          # ./generate_mov_entries.py hsw
+- name: movabs                                                           # ./generate_mov_entries.py hsw
+  operands:                                                              # ./generate_mov_entries.py hsw
+  - class: immediate                                                     # ./generate_mov_entries.py hsw
+    imd: int                                                             # ./generate_mov_entries.py hsw
+  - class: register                                                      # ./generate_mov_entries.py hsw
+    name: gpr                                                            # ./generate_mov_entries.py hsw
+  latency: 1                                                             # ./generate_mov_entries.py hsw
+  port_pressure: [[1, '0156']]                                           # ./generate_mov_entries.py hsw
+  throughput: 0.25                                                       # ./generate_mov_entries.py hsw
+  uops: 1                                                                # ./generate_mov_entries.py hsw
+                                                                         # ./generate_mov_entries.py hsw
 - name: mov  # with store, simple AGU                                    # ./generate_mov_entries.py hsw
   operands:                                                              # ./generate_mov_entries.py hsw
   - class: immediate                                                     # ./generate_mov_entries.py hsw

osaca/data/isa/x86.yml

@@ -353,6 +353,91 @@ instruction_forms:
           scale: "*"
           source: true
           destination: true
+    - name: cld
+      operands: []
+      hidden_operands:
+        - class: "flag"
+          name: "DF"
+          source: false
+          destination: true
+    - name: cbtw
+      operands: []
+      hidden_operands:
+        - class: "register"
+          name: "al"
+          source: true
+          destination: false
+        - class: "register"
+          name: "ax"
+          source: false
+          destination: true
+    - name: cwtl
+      operands: []
+      hidden_operands:
+        - class: "register"
+          name: "ax"
+          source: true
+          destination: false
+        - class: "register"
+          name: "eax"
+          source: false
+          destination: true
+    - name: cwtd
+      operands: []
+      hidden_operands:
+        - class: "register"
+          name: "ax"
+          source: true
+          destination: false
+        - class: "register"
+          name: "dx"
+          source: false
+          destination: true
+        - class: "register"
+          name: "ax"
+          source: false
+          destination: true
+    - name: cltd
+      operands: []
+      hidden_operands:
+        - class: "register"
+          name: "eax"
+          source: true
+          destination: false
+        - class: "register"
+          name: "edx"
+          source: false
+          destination: true
+        - class: "register"
+          name: "eax"
+          source: false
+          destination: true
+    - name: cltq
+      operands: []
+      hidden_operands:
+        - class: "register"
+          name: "eax"
+          source: true
+          destination: false
+        - class: "register"
+          name: "rax"
+          source: false
+          destination: true
+    - name: cqtq
+      operands: []
+      hidden_operands:
+        - class: "register"
+          name: "rax"
+          source: true
+          destination: false
+        - class: "register"
+          name: "rdx"
+          source: false
+          destination: true
+        - class: "register"
+          name: "rax"
+          source: false
+          destination: true
     - name: cmova
       operands:
         - class: "register"
@@ -2197,6 +2282,69 @@ instruction_forms:
           name: "PF"
           source: false
           destination: true
+    - name: cmpxchg
+      operands:
+        - class: "register"
+          name: "gpr"
+          source: true
+          destination: false
+        - class: "register"
+          name: "gpr"
+          source: true
+          destination: true
+      hidden_operands:
+        - class: "register"
+          name: "rax"
+          source: true
+          destination: true
+        - class: "flag"
+          name: "ZF"
+          source: true
+          destination: true
+    - name: cmpxchg
+      operands:
+        - class: "memory"
+          base: "*"
+          offset: "*"
+          index: "*"
+          scale: "*"
+          source: true
+          destination: false
+        - class: "register"
+          name: "gpr"
+          source: true
+          destination: true
+      hidden_operands:
+        - class: "register"
+          name: "rax"
+          source: true
+          destination: true
+        - class: "flag"
+          name: "ZF"
+          source: true
+          destination: true
+    - name: cmpxchg
+      operands:
+        - class: "register"
+          name: "gpr"
+          source: true
+          destination: false
+        - class: "memory"
+          base: "*"
+          offset: "*"
+          index: "*"
+          scale: "*"
+          source: true
+          destination: true
+      hidden_operands:
+        - class: "register"
+          name: "rax"
+          source: true
+          destination: true
+        - class: "flag"
+          name: "ZF"
+          source: true
+          destination: true
     - name: dec
       operands:
         - class: "register"

osaca/data/ivb.yml

@@ -175,6 +175,62 @@ instruction_forms:
   port_pressure: [[2, '015']]
   throughput: 0.6666666666666666
   uops: 3
+- name: cbtw  # AT&T version of CBW
+  operands: []
+  latency: 1
+  port_pressure: [[1, '015']]
+  throughput: 0.3333333333333333
+  uops: 1
+- name: cwtl  # AT&T version of CWDE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '015']]
+  throughput: 0.3333333333333333
+  uops: 1
+- name: cwtd  # AT&T version of CWD
+  operands: []
+  latency: 2
+  port_pressure: [[1, '015'], [1, '05']]
+  throughput: 0.8333333333333333
+  uops: 2
+- name: cltd  # AT&T version of CDQ
+  operands: []
+  latency: 1
+  port_pressure: [[1, '05']]
+  throughput: 0.5
+  uops: 1
+- name: cltq  # AT&T version of CDQE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '015']]
+  throughput: 0.3333333333333333
+  uops: 1
+- name: cqtq  # AT&T version of CQO
+  operands: []
+  latency: 1
+  port_pressure: [[1, '05']]
+  throughput: 0.5
+  uops: 1
+- name: [cmpeqsd, cmpltsd, cmplesd, cmpunordsd, cmpneqsd, cmpnltsd, cmpnlesd, cmpordsd]
+  # same as CMPSD xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
+- name: [cmpeqss, cmpltss, cmpless, cmpunordss, cmpneqss, cmpnltss, cmpnless, cmpordss]
+  # same as CMPSS xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
 - name: cmpps
   operands:
   - class: immediate
@@ -359,6 +415,38 @@ instruction_forms:
   port_pressure: [[1, '1']]  # JH: according to uops.info
   throughput: 1.0  # JH: measured on phinally
   uops: 1
+- name: [vcmpeqsd, vcmpltsd, vcmplesd, vcmpunordsd, vcmpneqsd, vcmpnltsd, vcmpnlesd, vcmpordsd,
+         vcmpeq_uqsd, vcmpngesd, vcmpngtsd, vcmpfalsesd, vcmpneq_oqsd, vcmpgesd, vcmpgtsd,
+         vcmptruesd, vcmpeq_ossd, vcmplt_oqsd, vcmple_oqsd, vcmpunord_ssd, vcmpneq_ussd,
+         vcmpnlt_uqsd, vcmpnle_uqsd, vcmpord_ssd, vcmpeq_ussd, vcmpnge_uqsd, vcmpngt_uqsd,
+         vcmpfalse_ossd, vcmpneq_ossd, vcmpge_oqsd, vcmpgt_oqsd, vcmptrue_ussd]
+  # same as VCMPSD xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
+- name: [vcmpeqss, vcmpltss, vcmpless, vcmpunordss, vcmpneqss, vcmpnltss, vcmpnless, vcmpordss,
+         vcmpeq_uqss, vcmpngess, vcmpngtss, vcmpfalsess, vcmpneq_oqss, vcmpgess, vcmpgtss,
+         vcmptruess, vcmpeq_osss, vcmplt_oqss, vcmple_oqss, vcmpunord_sss, vcmpneq_usss,
+         vcmpnlt_uqss, vcmpnle_uqss, vcmpord_sss, vcmpeq_usss, vcmpnge_uqss, vcmpngt_uqss,
+         vcmpfalse_osss, vcmpneq_osss, vcmpge_oqss, vcmpgt_oqss, vcmptrue_usss]
+  # same as VCMPSS xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
 - name: vcvtss2sd
   operands:
   - class: register
@@ -21001,6 +21089,17 @@ instruction_forms:
   throughput: 0.0                                                       # ./generate_mov_entries.py ivb
   uops: 0                                                               # ./generate_mov_entries.py ivb
                                                                         # ./generate_mov_entries.py ivb
+- name: movabs                                                          # ./generate_mov_entries.py ivb
+  operands:                                                             # ./generate_mov_entries.py ivb
+  - class: immediate                                                    # ./generate_mov_entries.py ivb
+    imd: int                                                            # ./generate_mov_entries.py ivb
+  - class: register                                                     # ./generate_mov_entries.py ivb
+    name: gpr                                                           # ./generate_mov_entries.py ivb
+  latency: 0                                                            # ./generate_mov_entries.py ivb
+  port_pressure: []                                                     # ./generate_mov_entries.py ivb
+  throughput: 0.0                                                       # ./generate_mov_entries.py ivb
+  uops: 0                                                               # ./generate_mov_entries.py ivb
+                                                                        # ./generate_mov_entries.py ivb
 - name: mov  # with store                                               # ./generate_mov_entries.py ivb
   operands:                                                             # ./generate_mov_entries.py ivb
   - class: immediate                                                    # ./generate_mov_entries.py ivb

osaca/data/snb.yml

@@ -62,6 +62,62 @@ instruction_forms:
   port_pressure: []
   throughput: 1
   uops: 0
+- name: cbtw  # AT&T version of CBW
+  operands: []
+  latency: 1
+  port_pressure: [[1, '015']]
+  throughput: 0.3333333333333333
+  uops: 1
+- name: cwtl  # AT&T version of CWDE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '015']]
+  throughput: 0.3333333333333333
+  uops: 1
+- name: cwtd  # AT&T version of CWD
+  operands: []
+  latency: 2
+  port_pressure: [[1, '015'], [1, '05']]
+  throughput: 0.8333333333333333
+  uops: 2
+- name: cltd  # AT&T version of CDQ
+  operands: []
+  latency: 1
+  port_pressure: [[1, '05']]
+  throughput: 0.5
+  uops: 1
+- name: cltq  # AT&T version of CDQE
+  operands: []
+  latency: 1
+  port_pressure: [[1, '015']]
+  throughput: 0.3333333333333333
+  uops: 1
+- name: cqtq  # AT&T version of CQO
+  operands: []
+  latency: 1
+  port_pressure: [[1, '05']]
+  throughput: 0.5
+  uops: 1
+- name: [cmpeqsd, cmpltsd, cmplesd, cmpunordsd, cmpneqsd, cmpnltsd, cmpnlesd, cmpordsd]
+  # same as CMPSD xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
+- name: [cmpeqss, cmpltss, cmpless, cmpunordss, cmpneqss, cmpnltss, cmpnless, cmpordss]
+  # same as CMPSS xmm1, xmm2, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
 - name: cmpeqps  # same as CMPPS xmm1, xmm2, 0
   operands:
   - class: register
@@ -222,6 +278,38 @@ instruction_forms:
   port_pressure: [[1, '1']]  # JH: according to uops.info
   throughput: 1.0  # JH: measured on phinally
   uops: 1
+- name: [vcmpeqsd, vcmpltsd, vcmplesd, vcmpunordsd, vcmpneqsd, vcmpnltsd, vcmpnlesd, vcmpordsd,
+         vcmpeq_uqsd, vcmpngesd, vcmpngtsd, vcmpfalsesd, vcmpneq_oqsd, vcmpgesd, vcmpgtsd,
+         vcmptruesd, vcmpeq_ossd, vcmplt_oqsd, vcmple_oqsd, vcmpunord_ssd, vcmpneq_ussd,
+         vcmpnlt_uqsd, vcmpnle_uqsd, vcmpord_ssd, vcmpeq_ussd, vcmpnge_uqsd, vcmpngt_uqsd,
+         vcmpfalse_ossd, vcmpneq_ossd, vcmpge_oqsd, vcmpgt_oqsd, vcmptrue_ussd]
+  # same as VCMPSD xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
+- name: [vcmpeqss, vcmpltss, vcmpless, vcmpunordss, vcmpneqss, vcmpnltss, vcmpnless, vcmpordss,
+         vcmpeq_uqss, vcmpngess, vcmpngtss, vcmpfalsess, vcmpneq_oqss, vcmpgess, vcmpgtss,
+         vcmptruess, vcmpeq_osss, vcmplt_oqss, vcmple_oqss, vcmpunord_sss, vcmpneq_usss,
+         vcmpnlt_uqss, vcmpnle_uqss, vcmpord_sss, vcmpeq_usss, vcmpnge_uqss, vcmpngt_uqss,
+         vcmpfalse_osss, vcmpneq_osss, vcmpge_oqss, vcmpgt_oqss, vcmptrue_usss]
+  # same as VCMPSS xmm, xmm, xmm, imd
+  operands:
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  - class: register
+    name: xmm
+  latency: 3
+  port_pressure: [[1, '1']]
+  throughput: 1
 - name: vpsrld
   operands:
   - class: register
@@ -20722,6 +20810,18 @@ instruction_forms:
   uops: 2                                                               # ./generate_mov_entries.py snb
                                                                         # ./generate_mov_entries.py snb
 - name: mov                                                             # ./generate_mov_entries.py snb
+  operands:                                                             # ./generate_mov_entries.py snb
+  - class: immediate                                                    # ./generate_mov_entries.py snb
+    imd: int                                                            # ./generate_mov_entries.py snb
+  - class: register                                                     # ./generate_mov_entries.py snb
+    name: gpr                                                           # ./generate_mov_entries.py snb
+  latency: 1                                                            # ./generate_mov_entries.py snb
+  port_pressure: [[1, '015']]                                           # ./generate_mov_entries.py snb
+  throughput: 0.3333333333333333                                        # ./generate_mov_entries.py snb
+  uops: 1                                                               # ./generate_mov_entries.py snb
+
+                                                                        # ./generate_mov_entries.py snb
+- name: movabs                                                          # ./generate_mov_entries.py snb
   operands:                                                             # ./generate_mov_entries.py snb
   - class: immediate                                                    # ./generate_mov_entries.py snb
     imd: int                                                            # ./generate_mov_entries.py snb