VZEROUPPER - Port Usage (IACA 3.0)


With blocking instructions for port '0':

Throughput Analysis Report
--------------------------
Block Throughput: 9.95 Cycles       Throughput Bottleneck: Backend
Loop Count:  22
Port Binding In Cycles Per Iteration:
--------------------------------------------------------------------------------------------------
|  Port  |   0   -  DV   |   1   |   2   -  D    |   3   -  D    |   4   |   5   |   6   |   7   |
--------------------------------------------------------------------------------------------------
| Cycles | 10.0     0.0  |  1.3  |  0.0     0.0  |  0.0     0.0  |  0.0  |  1.4  |  1.3  |  0.0  |
--------------------------------------------------------------------------------------------------

| Num Of   |                    Ports pressure in cycles                         |      |
|  Uops    |  0  - DV    |  1   |  2  -  D    |  3  -  D    |  4   |  5   |  6   |  7   |
-----------------------------------------------------------------------------------------
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm0, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm2, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm3, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm4, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm5, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm6, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm7, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm8, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm9, xmm1
|   2      | 1.0         |      |             |             |      |      |      |      | vcvtph2ps xmm10, xmm1
|   4      |             | 1.3  |             |             |      | 1.4  | 1.3  |      | vzeroupper 
Total Num Of Uops: 24

With blocking instructions for port '1':

Throughput Analysis Report
--------------------------
Block Throughput: 9.95 Cycles       Throughput Bottleneck: Dependency chains
Loop Count:  23
Port Binding In Cycles Per Iteration:
--------------------------------------------------------------------------------------------------
|  Port  |   0   -  DV   |   1   |   2   -  D    |   3   -  D    |   4   |   5   |   6   |   7   |
--------------------------------------------------------------------------------------------------
| Cycles |  1.3     0.0  | 10.0  |  0.0     0.0  |  0.0     0.0  |  0.0  |  1.4  |  1.4  |  0.0  |
--------------------------------------------------------------------------------------------------

| Num Of   |                    Ports pressure in cycles                         |      |
|  Uops    |  0  - DV    |  1   |  2  -  D    |  3  -  D    |  4   |  5   |  6   |  7   |
-----------------------------------------------------------------------------------------
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm0, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm3, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm4, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm5, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm6, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm7, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm8, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm9, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm10, xmm1, xmm2
|   3      |             | 1.0  |             |             |      |      |      |      | vhaddpd xmm11, xmm1, xmm2
|   4      | 1.3         |      |             |             |      | 1.4  | 1.4  |      | vzeroupper 
Total Num Of Uops: 34

With blocking instructions for port '5':

Throughput Analysis Report
--------------------------
Block Throughput: 9.95 Cycles       Throughput Bottleneck: Backend
Loop Count:  22
Port Binding In Cycles Per Iteration:
--------------------------------------------------------------------------------------------------
|  Port  |   0   -  DV   |   1   |   2   -  D    |   3   -  D    |   4   |   5   |   6   |   7   |
--------------------------------------------------------------------------------------------------
| Cycles |  1.3     0.0  |  1.4  |  0.0     0.0  |  0.0     0.0  |  0.0  | 10.0  |  1.3  |  0.0  |
--------------------------------------------------------------------------------------------------

| Num Of   |                    Ports pressure in cycles                         |      |
|  Uops    |  0  - DV    |  1   |  2  -  D    |  3  -  D    |  4   |  5   |  6   |  7   |
-----------------------------------------------------------------------------------------
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx0, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx1, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx2, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx3, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx4, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx5, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx6, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx7, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx0, r8d
|   1      |             |      |             |             |      | 1.0  |      |      | movd mmx1, r8d
|   4      | 1.3         | 1.4  |             |             |      |      | 1.3  |      | vzeroupper 
Total Num Of Uops: 14

With blocking instructions for ports {'0', '1'}:

Throughput Analysis Report
--------------------------
Block Throughput: 4.95 Cycles       Throughput Bottleneck: Backend
Loop Count:  22
Port Binding In Cycles Per Iteration:
--------------------------------------------------------------------------------------------------
|  Port  |   0   -  DV   |   1   |   2   -  D    |   3   -  D    |   4   |   5   |   6   |   7   |
--------------------------------------------------------------------------------------------------
| Cycles |  5.0     0.0  |  5.0  |  0.0     0.0  |  0.0     0.0  |  0.0  |  2.0  |  2.0  |  0.0  |
--------------------------------------------------------------------------------------------------

| Num Of   |                    Ports pressure in cycles                         |      |
|  Uops    |  0  - DV    |  1   |  2  -  D    |  3  -  D    |  4   |  5   |  6   |  7   |
-----------------------------------------------------------------------------------------
|   1      | 1.0         |      |             |             |      |      |      |      | vblendpd xmm0, xmm1, xmm2, 0x2
|   1      |             | 1.0  |             |             |      |      |      |      | vblendpd xmm3, xmm1, xmm2, 0x2
|   1      | 1.0         |      |             |             |      |      |      |      | vblendpd xmm4, xmm1, xmm2, 0x2
|   1      |             | 1.0  |             |             |      |      |      |      | vblendpd xmm5, xmm1, xmm2, 0x2
|   1      | 1.0         |      |             |             |      |      |      |      | vblendpd xmm6, xmm1, xmm2, 0x2
|   1      |             | 1.0  |             |             |      |      |      |      | vblendpd xmm7, xmm1, xmm2, 0x2
|   1      | 1.0         |      |             |             |      |      |      |      | vblendpd xmm8, xmm1, xmm2, 0x2
|   1      |             | 1.0  |             |             |      |      |      |      | vblendpd xmm9, xmm1, xmm2, 0x2
|   1      | 1.0         |      |             |             |      |      |      |      | vblendpd xmm10, xmm1, xmm2, 0x2
|   1      |             | 1.0  |             |             |      |      |      |      | vblendpd xmm11, xmm1, xmm2, 0x2
|   4      |             |      |             |             |      | 2.0  | 2.0  |      | vzeroupper 
Total Num Of Uops: 14

With blocking instructions for ports {'0', '6'}:

Throughput Analysis Report
--------------------------
Block Throughput: 4.95 Cycles       Throughput Bottleneck: Backend
Loop Count:  22
Port Binding In Cycles Per Iteration:
--------------------------------------------------------------------------------------------------
|  Port  |   0   -  DV   |   1   |   2   -  D    |   3   -  D    |   4   |   5   |   6   |   7   |
--------------------------------------------------------------------------------------------------
| Cycles |  5.0     0.0  |  2.0  |  0.0     0.0  |  0.0     0.0  |  0.0  |  2.0  |  5.0  |  0.0  |
--------------------------------------------------------------------------------------------------

| Num Of   |                    Ports pressure in cycles                         |      |
|  Uops    |  0  - DV    |  1   |  2  -  D    |  3  -  D    |  4   |  5   |  6   |  7   |
-----------------------------------------------------------------------------------------
|   1      | 1.0         |      |             |             |      |      |      |      | rorx r8d, r9d, 0x2
|   1      |             |      |             |             |      |      | 1.0  |      | rorx r10d, r9d, 0x2
|   1      | 1.0         |      |             |             |      |      |      |      | rorx r11d, r9d, 0x2
|   1      |             |      |             |             |      |      | 1.0  |      | rorx r12d, r9d, 0x2
|   1      | 1.0         |      |             |             |      |      |      |      | rorx ebx, r9d, 0x2
|   1      |             |      |             |             |      |      | 1.0  |      | rorx ecx, r9d, 0x2
|   1      | 1.0         |      |             |             |      |      |      |      | rorx edx, r9d, 0x2
|   1      |             |      |             |             |      |      | 1.0  |      | rorx r8d, r9d, 0x2
|   1      | 1.0         |      |             |             |      |      |      |      | rorx r10d, r9d, 0x2
|   1      |             |      |             |             |      |      | 1.0  |      | rorx r11d, r9d, 0x2
|   4      |             | 2.0  |             |             |      | 2.0  |      |      | vzeroupper 
Total Num Of Uops: 14

With blocking instructions for ports {'1', '5'}:

Throughput Analysis Report
--------------------------
Block Throughput: 4.95 Cycles       Throughput Bottleneck: Backend
Loop Count:  22
Port Binding In Cycles Per Iteration:
--------------------------------------------------------------------------------------------------
|  Port  |   0   -  DV   |   1   |   2   -  D    |   3   -  D    |   4   |   5   |   6   |   7   |
--------------------------------------------------------------------------------------------------
| Cycles |  2.0     0.0  |  5.0  |  0.0     0.0  |  0.0     0.0  |  0.0  |  5.0  |  2.0  |  0.0  |
--------------------------------------------------------------------------------------------------

| Num Of   |                    Ports pressure in cycles                         |      |
|  Uops    |  0  - DV    |  1   |  2  -  D    |  3  -  D    |  4   |  5   |  6   |  7   |
-----------------------------------------------------------------------------------------
|   1      |             | 1.0  |             |             |      |      |      |      | lea r8w, ptr [r14]
|   1      |             |      |             |             |      | 1.0  |      |      | lea r9w, ptr [r14]
|   1      |             | 1.0  |             |             |      |      |      |      | lea r10w, ptr [r14]
|   1      |             |      |             |             |      | 1.0  |      |      | lea r11w, ptr [r14]
|   1      |             | 1.0  |             |             |      |      |      |      | lea r12w, ptr [r14]
|   1      |             |      |             |             |      | 1.0  |      |      | lea bx, ptr [r14]
|   1      |             | 1.0  |             |             |      |      |      |      | lea cx, ptr [r14]
|   1      |             |      |             |             |      | 1.0  |      |      | lea dx, ptr [r14]
|   1      |             | 1.0  |             |             |      |      |      |      | lea r8w, ptr [r14]
|   1      |             |      |             |             |      | 1.0  |      |      | lea r9w, ptr [r14]
|   4      | 2.0         |      |             |             |      |      | 2.0  |      | vzeroupper 
Total Num Of Uops: 14

With blocking instructions for ports {'0', '1', '5'}:

Throughput Analysis Report
--------------------------
Block Throughput: 4.29 Cycles       Throughput Bottleneck: FrontEnd
Loop Count:  24
Port Binding In Cycles Per Iteration:
--------------------------------------------------------------------------------------------------
|  Port  |   0   -  DV   |   1   |   2   -  D    |   3   -  D    |   4   |   5   |   6   |   7   |
--------------------------------------------------------------------------------------------------
| Cycles |  4.3     0.0  |  4.3  |  0.0     0.0  |  0.0     0.0  |  0.0  |  4.3  |  3.0  |  0.0  |
--------------------------------------------------------------------------------------------------

| Num Of   |                    Ports pressure in cycles                         |      |
|  Uops    |  0  - DV    |  1   |  2  -  D    |  3  -  D    |  4   |  5   |  6   |  7   |
-----------------------------------------------------------------------------------------
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm0, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm2, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm3, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm4, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm5, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm6, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm7, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm8, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm9, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm10, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm11, xmm1
|   1      | 0.3         | 0.3  |             |             |      | 0.3  |      |      | vmovq xmm12, xmm1
|   4      | 0.3         | 0.3  |             |             |      | 0.3  | 3.0  |      | vzeroupper 
Total Num Of Uops: 16
⇨ One μop that can only use ports {'0', '1', '5'}

With blocking instructions for ports {'0', '1', '5', '6'}:

Throughput Analysis Report
--------------------------
Block Throughput: 4.95 Cycles       Throughput Bottleneck: Dependency chains
Loop Count:  22
Port Binding In Cycles Per Iteration:
--------------------------------------------------------------------------------------------------
|  Port  |   0   -  DV   |   1   |   2   -  D    |   3   -  D    |   4   |   5   |   6   |   7   |
--------------------------------------------------------------------------------------------------
| Cycles |  5.0     0.0  |  5.0  |  0.0     0.0  |  0.0     0.0  |  0.0  |  5.0  |  5.0  |  0.0  |
--------------------------------------------------------------------------------------------------

| Num Of   |                    Ports pressure in cycles                         |      |
|  Uops    |  0  - DV    |  1   |  2  -  D    |  3  -  D    |  4   |  5   |  6   |  7   |
-----------------------------------------------------------------------------------------
|   1      | 1.0         |      |             |             |      |      |      |      | mov bh, 0x0
|   1      |             | 1.0  |             |             |      |      |      |      | mov ch, 0x0
|   1      |             |      |             |             |      | 1.0  |      |      | mov dh, 0x0
|   1      |             |      |             |             |      |      | 1.0  |      | mov ah, 0x0
|   1      | 1.0         |      |             |             |      |      |      |      | mov bh, 0x0
|   1      |             | 1.0  |             |             |      |      |      |      | mov ch, 0x0
|   1      |             |      |             |             |      | 1.0  |      |      | mov dh, 0x0
|   1      |             |      |             |             |      |      | 1.0  |      | mov ah, 0x0
|   1      | 1.0         |      |             |             |      |      |      |      | mov bh, 0x0
|   1      |             | 1.0  |             |             |      |      |      |      | mov ch, 0x0
|   1      |             |      |             |             |      | 1.0  |      |      | mov dh, 0x0
|   1      |             |      |             |             |      |      | 1.0  |      | mov ah, 0x0
|   1      | 1.0         |      |             |             |      |      |      |      | mov bh, 0x0
|   1      |             | 1.0  |             |             |      |      |      |      | mov ch, 0x0
|   1      |             |      |             |             |      | 1.0  |      |      | mov dh, 0x0
|   1      |             |      |             |             |      |      | 1.0  |      | mov ah, 0x0
|   4      | 1.0         | 1.0  |             |             |      | 1.0  | 1.0  |      | vzeroupper 
Total Num Of Uops: 20
⇨ 3 μops that can only use ports {'0', '1', '5', '6'}