Skip to content

[AMDGPU][MC][True16] Support V_RCP/SQRT/RSQ/LOG/EXP_F16. #81131

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Feb 19, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions llvm/lib/Target/AMDGPU/VOP1Instructions.td
Original file line number Diff line number Diff line change
Expand Up @@ -954,10 +954,15 @@ defm V_CVT_F16_U16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x050, "v_cvt_f16_
defm V_CVT_F16_I16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x051, "v_cvt_f16_i16">;
defm V_CVT_U16_F16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x052, "v_cvt_u16_f16">;
defm V_CVT_I16_F16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x053, "v_cvt_i16_f16">;
defm V_RCP_F16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x054, "v_rcp_f16">;
defm V_RCP_F16_fake16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x054, "v_rcp_f16">;
defm V_SQRT_F16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x055, "v_sqrt_f16">;
defm V_SQRT_F16_fake16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x055, "v_sqrt_f16">;
defm V_RSQ_F16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x056, "v_rsq_f16">;
defm V_RSQ_F16_fake16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x056, "v_rsq_f16">;
defm V_LOG_F16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x057, "v_log_f16">;
defm V_LOG_F16_fake16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x057, "v_log_f16">;
defm V_EXP_F16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x058, "v_exp_f16">;
defm V_EXP_F16_fake16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x058, "v_exp_f16">;
defm V_FREXP_MANT_F16_fake16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x059, "v_frexp_mant_f16">;
defm V_FREXP_EXP_I16_F16_t16 : VOP1_Real_FULL_t16_gfx11_gfx12<0x05a, "v_frexp_exp_i16_f16">;
Expand Down
222 changes: 222 additions & 0 deletions llvm/test/MC/AMDGPU/gfx11_asm_vop1-fake16.s
Original file line number Diff line number Diff line change
Expand Up @@ -93,3 +93,225 @@ v_ceil_f16 v5, src_scc

v_ceil_f16 v127, 0xfe0b
// GFX11: encoding: [0xff,0xb8,0xfe,0x7e,0x0b,0xfe,0x00,0x00]

v_rcp_f16 v5, v1
// GFX11: encoding: [0x01,0xa9,0x0a,0x7e]

v_rcp_f16 v5, v127
// GFX11: encoding: [0x7f,0xa9,0x0a,0x7e]

v_rcp_f16 v5, s1
// GFX11: encoding: [0x01,0xa8,0x0a,0x7e]

v_rcp_f16 v5, s105
// GFX11: encoding: [0x69,0xa8,0x0a,0x7e]

v_rcp_f16 v5, vcc_lo
// GFX11: encoding: [0x6a,0xa8,0x0a,0x7e]

v_rcp_f16 v5, vcc_hi
// GFX11: encoding: [0x6b,0xa8,0x0a,0x7e]

v_rcp_f16 v5, ttmp15
// GFX11: encoding: [0x7b,0xa8,0x0a,0x7e]

v_rcp_f16 v5, m0
// GFX11: encoding: [0x7d,0xa8,0x0a,0x7e]

v_rcp_f16 v5, exec_lo
// GFX11: encoding: [0x7e,0xa8,0x0a,0x7e]

v_rcp_f16 v5, exec_hi
// GFX11: encoding: [0x7f,0xa8,0x0a,0x7e]

v_rcp_f16 v5, null
// GFX11: encoding: [0x7c,0xa8,0x0a,0x7e]

v_rcp_f16 v5, -1
// GFX11: encoding: [0xc1,0xa8,0x0a,0x7e]

v_rcp_f16 v5, 0.5
// GFX11: encoding: [0xf0,0xa8,0x0a,0x7e]

v_rcp_f16 v5, src_scc
// GFX11: encoding: [0xfd,0xa8,0x0a,0x7e]

v_rcp_f16 v127, 0xfe0b
// GFX11: encoding: [0xff,0xa8,0xfe,0x7e,0x0b,0xfe,0x00,0x00]

v_sqrt_f16 v5, v1
// GFX11: encoding: [0x01,0xab,0x0a,0x7e]

v_sqrt_f16 v5, v127
// GFX11: encoding: [0x7f,0xab,0x0a,0x7e]

v_sqrt_f16 v5, s1
// GFX11: encoding: [0x01,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, s105
// GFX11: encoding: [0x69,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, vcc_lo
// GFX11: encoding: [0x6a,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, vcc_hi
// GFX11: encoding: [0x6b,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, ttmp15
// GFX11: encoding: [0x7b,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, m0
// GFX11: encoding: [0x7d,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, exec_lo
// GFX11: encoding: [0x7e,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, exec_hi
// GFX11: encoding: [0x7f,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, null
// GFX11: encoding: [0x7c,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, -1
// GFX11: encoding: [0xc1,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, 0.5
// GFX11: encoding: [0xf0,0xaa,0x0a,0x7e]

v_sqrt_f16 v5, src_scc
// GFX11: encoding: [0xfd,0xaa,0x0a,0x7e]

v_sqrt_f16 v127, 0xfe0b
// GFX11: encoding: [0xff,0xaa,0xfe,0x7e,0x0b,0xfe,0x00,0x00]

v_rsq_f16 v5, v1
// GFX11: encoding: [0x01,0xad,0x0a,0x7e]

v_rsq_f16 v5, v127
// GFX11: encoding: [0x7f,0xad,0x0a,0x7e]

v_rsq_f16 v5, s1
// GFX11: encoding: [0x01,0xac,0x0a,0x7e]

v_rsq_f16 v5, s105
// GFX11: encoding: [0x69,0xac,0x0a,0x7e]

v_rsq_f16 v5, vcc_lo
// GFX11: encoding: [0x6a,0xac,0x0a,0x7e]

v_rsq_f16 v5, vcc_hi
// GFX11: encoding: [0x6b,0xac,0x0a,0x7e]

v_rsq_f16 v5, ttmp15
// GFX11: encoding: [0x7b,0xac,0x0a,0x7e]

v_rsq_f16 v5, m0
// GFX11: encoding: [0x7d,0xac,0x0a,0x7e]

v_rsq_f16 v5, exec_lo
// GFX11: encoding: [0x7e,0xac,0x0a,0x7e]

v_rsq_f16 v5, exec_hi
// GFX11: encoding: [0x7f,0xac,0x0a,0x7e]

v_rsq_f16 v5, null
// GFX11: encoding: [0x7c,0xac,0x0a,0x7e]

v_rsq_f16 v5, -1
// GFX11: encoding: [0xc1,0xac,0x0a,0x7e]

v_rsq_f16 v5, 0.5
// GFX11: encoding: [0xf0,0xac,0x0a,0x7e]

v_rsq_f16 v5, src_scc
// GFX11: encoding: [0xfd,0xac,0x0a,0x7e]

v_log_f16 v5, v1
// GFX11: encoding: [0x01,0xaf,0x0a,0x7e]

v_log_f16 v5, v127
// GFX11: encoding: [0x7f,0xaf,0x0a,0x7e]

v_log_f16 v5, s1
// GFX11: encoding: [0x01,0xae,0x0a,0x7e]

v_log_f16 v5, s105
// GFX11: encoding: [0x69,0xae,0x0a,0x7e]

v_log_f16 v5, vcc_lo
// GFX11: encoding: [0x6a,0xae,0x0a,0x7e]

v_log_f16 v5, vcc_hi
// GFX11: encoding: [0x6b,0xae,0x0a,0x7e]

v_log_f16 v5, ttmp15
// GFX11: encoding: [0x7b,0xae,0x0a,0x7e]

v_log_f16 v5, m0
// GFX11: encoding: [0x7d,0xae,0x0a,0x7e]

v_log_f16 v5, exec_lo
// GFX11: encoding: [0x7e,0xae,0x0a,0x7e]

v_log_f16 v5, exec_hi
// GFX11: encoding: [0x7f,0xae,0x0a,0x7e]

v_log_f16 v5, null
// GFX11: encoding: [0x7c,0xae,0x0a,0x7e]

v_log_f16 v5, -1
// GFX11: encoding: [0xc1,0xae,0x0a,0x7e]

v_log_f16 v5, 0.5
// GFX11: encoding: [0xf0,0xae,0x0a,0x7e]

v_log_f16 v5, src_scc
// GFX11: encoding: [0xfd,0xae,0x0a,0x7e]

v_log_f16 v127, 0xfe0b
// GFX11: encoding: [0xff,0xae,0xfe,0x7e,0x0b,0xfe,0x00,0x00]

v_exp_f16 v5, v1
// GFX11: encoding: [0x01,0xb1,0x0a,0x7e]

v_exp_f16 v5, v127
// GFX11: encoding: [0x7f,0xb1,0x0a,0x7e]

v_exp_f16 v5, s1
// GFX11: encoding: [0x01,0xb0,0x0a,0x7e]

v_exp_f16 v5, s105
// GFX11: encoding: [0x69,0xb0,0x0a,0x7e]

v_exp_f16 v5, vcc_lo
// GFX11: encoding: [0x6a,0xb0,0x0a,0x7e]

v_exp_f16 v5, vcc_hi
// GFX11: encoding: [0x6b,0xb0,0x0a,0x7e]

v_exp_f16 v5, ttmp15
// GFX11: encoding: [0x7b,0xb0,0x0a,0x7e]

v_exp_f16 v5, m0
// GFX11: encoding: [0x7d,0xb0,0x0a,0x7e]

v_exp_f16 v5, exec_lo
// GFX11: encoding: [0x7e,0xb0,0x0a,0x7e]

v_exp_f16 v5, exec_hi
// GFX11: encoding: [0x7f,0xb0,0x0a,0x7e]

v_exp_f16 v5, null
// GFX11: encoding: [0x7c,0xb0,0x0a,0x7e]

v_exp_f16 v5, -1
// GFX11: encoding: [0xc1,0xb0,0x0a,0x7e]

v_exp_f16 v5, 0.5
// GFX11: encoding: [0xf0,0xb0,0x0a,0x7e]

v_exp_f16 v5, src_scc
// GFX11: encoding: [0xfd,0xb0,0x0a,0x7e]

v_exp_f16 v127, 0xfe0b
// GFX11: encoding: [0xff,0xb0,0xfe,0x7e,0x0b,0xfe,0x00,0x00]
Loading