Skip to content

Commit c45fca8

Browse files
lukel97lukel97
committed
Precommit tests
1 parent fe56c8f commit c45fca8

File tree

5 files changed

+1557
-0
lines changed

5 files changed

+1557
-0
lines changed

‎llvm/test/CodeGen/RISCV/rvv/fixed-vectors-deinterleave-load.ll

Lines changed: 164 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -257,6 +257,49 @@ define {<2 x double>, <2 x double>} @vector_deinterleave_load_v2f64_v4f64(ptr %p
257257
ret {<2 x double>, <2 x double>} %res1
258258
}
259259

260+
define { <8 x i8>, <8 x i8>, <8 x i8> } @vector_deinterleave_load_factor3(ptr %p) {
261+
; CHECK-LABEL: vector_deinterleave_load_factor3:
262+
; CHECK: # %bb.0:
263+
; CHECK-NEXT: addi sp, sp, -16
264+
; CHECK-NEXT: .cfi_def_cfa_offset 16
265+
; CHECK-NEXT: csrr a1, vlenb
266+
; CHECK-NEXT: slli a1, a1, 1
267+
; CHECK-NEXT: sub sp, sp, a1
268+
; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 2 * vlenb
269+
; CHECK-NEXT: vsetivli zero, 24, e8, m2, ta, ma
270+
; CHECK-NEXT: vle8.v v8, (a0)
271+
; CHECK-NEXT: csrr a0, vlenb
272+
; CHECK-NEXT: srli a0, a0, 1
273+
; CHECK-NEXT: add a1, a0, a0
274+
; CHECK-NEXT: vsetivli zero, 8, e8, m1, ta, ma
275+
; CHECK-NEXT: vslidedown.vi v12, v8, 8
276+
; CHECK-NEXT: vsetivli zero, 8, e8, m2, ta, ma
277+
; CHECK-NEXT: vslidedown.vi v10, v8, 16
278+
; CHECK-NEXT: vsetvli zero, a1, e8, m1, ta, ma
279+
; CHECK-NEXT: vslideup.vx v8, v12, a0
280+
; CHECK-NEXT: addi a0, sp, 16
281+
; CHECK-NEXT: vmv1r.v v9, v10
282+
; CHECK-NEXT: vs2r.v v8, (a0)
283+
; CHECK-NEXT: vsetvli a1, zero, e8, mf2, ta, ma
284+
; CHECK-NEXT: vlseg3e8.v v6, (a0)
285+
; CHECK-NEXT: csrr a0, vlenb
286+
; CHECK-NEXT: slli a0, a0, 1
287+
; CHECK-NEXT: add sp, sp, a0
288+
; CHECK-NEXT: .cfi_def_cfa sp, 16
289+
; CHECK-NEXT: addi sp, sp, 16
290+
; CHECK-NEXT: .cfi_def_cfa_offset 0
291+
; CHECK-NEXT: ret
292+
%vec = load <24 x i8>, ptr %p
293+
%d0 = call {<8 x i8>, <8 x i8>, <8 x i8>} @llvm.vector.deinterleave3(<24 x i8> %vec)
294+
%t0 = extractvalue {<8 x i8>, <8 x i8>, <8 x i8>} %d0, 0
295+
%t1 = extractvalue {<8 x i8>, <8 x i8>, <8 x i8>} %d0, 1
296+
%t2 = extractvalue {<8 x i8>, <8 x i8>, <8 x i8>} %d0, 2
297+
%res0 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8> } poison, <8 x i8> %t0, 0
298+
%res1 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8> } %res0, <8 x i8> %t1, 0
299+
%res2 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8> } %res1, <8 x i8> %t2, 0
300+
ret { <8 x i8>, <8 x i8>, <8 x i8> } %res2
301+
}
302+
260303
define { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } @vector_deinterleave_load_factor4(ptr %p) {
261304
; CHECK-LABEL: vector_deinterleave_load_factor4:
262305
; CHECK: # %bb.0:
@@ -281,6 +324,127 @@ define { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } @vector_deinterleave_load_fact
281324
ret { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res3
282325
}
283326

327+
define { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } @vector_deinterleave_load_factor5(ptr %p) {
328+
; CHECK-LABEL: vector_deinterleave_load_factor5:
329+
; CHECK: # %bb.0:
330+
; CHECK-NEXT: addi sp, sp, -16
331+
; CHECK-NEXT: .cfi_def_cfa_offset 16
332+
; CHECK-NEXT: csrr a1, vlenb
333+
; CHECK-NEXT: slli a1, a1, 2
334+
; CHECK-NEXT: sub sp, sp, a1
335+
; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 4 * vlenb
336+
; CHECK-NEXT: li a1, 40
337+
; CHECK-NEXT: vsetvli zero, a1, e8, m4, ta, ma
338+
; CHECK-NEXT: vle8.v v8, (a0)
339+
; CHECK-NEXT: csrr a0, vlenb
340+
; CHECK-NEXT: srli a0, a0, 1
341+
; CHECK-NEXT: add a1, a0, a0
342+
; CHECK-NEXT: vsetivli zero, 8, e8, m2, ta, ma
343+
; CHECK-NEXT: vslidedown.vi v12, v8, 24
344+
; CHECK-NEXT: vslidedown.vi v14, v8, 16
345+
; CHECK-NEXT: vsetivli zero, 8, e8, m1, ta, ma
346+
; CHECK-NEXT: vslidedown.vi v13, v8, 8
347+
; CHECK-NEXT: vsetvli zero, a1, e8, m1, ta, ma
348+
; CHECK-NEXT: vslideup.vx v14, v12, a0
349+
; CHECK-NEXT: vmv1r.v v12, v8
350+
; CHECK-NEXT: vslideup.vx v12, v13, a0
351+
; CHECK-NEXT: li a0, 32
352+
; CHECK-NEXT: vsetivli zero, 8, e8, m4, ta, ma
353+
; CHECK-NEXT: vslidedown.vx v8, v8, a0
354+
; CHECK-NEXT: vmv1r.v v13, v14
355+
; CHECK-NEXT: addi a0, sp, 16
356+
; CHECK-NEXT: vmv2r.v v14, v8
357+
; CHECK-NEXT: vs4r.v v12, (a0)
358+
; CHECK-NEXT: vsetvli a1, zero, e8, mf2, ta, ma
359+
; CHECK-NEXT: vlseg5e8.v v8, (a0)
360+
; CHECK-NEXT: csrr a0, vlenb
361+
; CHECK-NEXT: slli a0, a0, 2
362+
; CHECK-NEXT: add sp, sp, a0
363+
; CHECK-NEXT: .cfi_def_cfa sp, 16
364+
; CHECK-NEXT: addi sp, sp, 16
365+
; CHECK-NEXT: .cfi_def_cfa_offset 0
366+
; CHECK-NEXT: ret
367+
%vec = load <40 x i8>, ptr %p
368+
%d0 = call {<8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>} @llvm.vector.deinterleave5(<40 x i8> %vec)
369+
%t0 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 0
370+
%t1 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 1
371+
%t2 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 2
372+
%t3 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 3
373+
%t4 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 4
374+
%res0 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } poison, <8 x i8> %t0, 0
375+
%res1 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res0, <8 x i8> %t1, 1
376+
%res2 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res1, <8 x i8> %t2, 2
377+
%res3 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res2, <8 x i8> %t3, 3
378+
%res4 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res3, <8 x i8> %t4, 4
379+
ret { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res4
380+
}
381+
382+
define { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } @vector_deinterleave_load_factor7(ptr %p) {
383+
; CHECK-LABEL: vector_deinterleave_load_factor7:
384+
; CHECK: # %bb.0:
385+
; CHECK-NEXT: addi sp, sp, -16
386+
; CHECK-NEXT: .cfi_def_cfa_offset 16
387+
; CHECK-NEXT: csrr a1, vlenb
388+
; CHECK-NEXT: slli a1, a1, 2
389+
; CHECK-NEXT: sub sp, sp, a1
390+
; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 4 * vlenb
391+
; CHECK-NEXT: li a1, 56
392+
; CHECK-NEXT: vsetvli zero, a1, e8, m4, ta, ma
393+
; CHECK-NEXT: vle8.v v8, (a0)
394+
; CHECK-NEXT: csrr a0, vlenb
395+
; CHECK-NEXT: li a1, 40
396+
; CHECK-NEXT: li a2, 32
397+
; CHECK-NEXT: vsetivli zero, 8, e8, m4, ta, ma
398+
; CHECK-NEXT: vslidedown.vx v16, v8, a1
399+
; CHECK-NEXT: li a1, 48
400+
; CHECK-NEXT: srli a0, a0, 1
401+
; CHECK-NEXT: vslidedown.vx v12, v8, a2
402+
; CHECK-NEXT: add a2, a0, a0
403+
; CHECK-NEXT: vsetivli zero, 8, e8, m2, ta, ma
404+
; CHECK-NEXT: vslidedown.vi v14, v8, 24
405+
; CHECK-NEXT: vslidedown.vi v18, v8, 16
406+
; CHECK-NEXT: vsetivli zero, 8, e8, m1, ta, ma
407+
; CHECK-NEXT: vslidedown.vi v13, v8, 8
408+
; CHECK-NEXT: vsetvli zero, a2, e8, m1, ta, ma
409+
; CHECK-NEXT: vslideup.vx v18, v14, a0
410+
; CHECK-NEXT: vsetivli zero, 8, e8, m4, ta, ma
411+
; CHECK-NEXT: vslidedown.vx v20, v8, a1
412+
; CHECK-NEXT: vsetvli zero, a2, e8, m1, ta, ma
413+
; CHECK-NEXT: vslideup.vx v8, v13, a0
414+
; CHECK-NEXT: vslideup.vx v12, v16, a0
415+
; CHECK-NEXT: vmv1r.v v9, v18
416+
; CHECK-NEXT: addi a0, sp, 16
417+
; CHECK-NEXT: vmv1r.v v13, v20
418+
; CHECK-NEXT: vmv2r.v v10, v12
419+
; CHECK-NEXT: vs4r.v v8, (a0)
420+
; CHECK-NEXT: vsetvli a1, zero, e8, mf2, ta, ma
421+
; CHECK-NEXT: vlseg7e8.v v8, (a0)
422+
; CHECK-NEXT: csrr a0, vlenb
423+
; CHECK-NEXT: slli a0, a0, 2
424+
; CHECK-NEXT: add sp, sp, a0
425+
; CHECK-NEXT: .cfi_def_cfa sp, 16
426+
; CHECK-NEXT: addi sp, sp, 16
427+
; CHECK-NEXT: .cfi_def_cfa_offset 0
428+
; CHECK-NEXT: ret
429+
%vec = load <56 x i8>, ptr %p
430+
%d0 = call {<8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>} @llvm.vector.deinterleave7(<56 x i8> %vec)
431+
%t0 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 0
432+
%t1 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 1
433+
%t2 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 2
434+
%t3 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 3
435+
%t4 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 4
436+
%t5 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 5
437+
%t6 = extractvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %d0, 6
438+
%res0 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } poison, <8 x i8> %t0, 0
439+
%res1 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res0, <8 x i8> %t1, 1
440+
%res2 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res1, <8 x i8> %t2, 2
441+
%res3 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res2, <8 x i8> %t3, 3
442+
%res4 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res3, <8 x i8> %t4, 4
443+
%res5 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res3, <8 x i8> %t5, 5
444+
%res6 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res3, <8 x i8> %t6, 6
445+
ret { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res6
446+
}
447+
284448
define {<2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>} @vector_deinterleave_load_factor8(ptr %ptr) {
285449
; CHECK-LABEL: vector_deinterleave_load_factor8:
286450
; CHECK: # %bb.0:

0 commit comments

Comments
 (0)