metal : support tensors in mul_mm_id

ggml/src/ggml-metal/ggml-metal.metal

@@ -8207,9 +8207,9 @@ kernel void kernel_mul_mm(
     auto tA = tensor<threadgroup S0, dextents<int32_t, 2>, tensor_inline>(sa, dextents<int32_t, 2>(NK,  NR0));
     auto tB = tensor<threadgroup S1, dextents<int32_t, 2>, tensor_inline>(sb, dextents<int32_t, 2>(NR1, NK ));
 
-    constexpr auto desc = mpp::tensor_ops::matmul2d_descriptor(NR1, NR0, NK, false, true, false, mpp::tensor_ops::matmul2d_descriptor::mode::multiply_accumulate);
+    mpp::tensor_ops::matmul2d<
-
+        mpp::tensor_ops::matmul2d_descriptor(NR1, NR0, NK, false, true, false, mpp::tensor_ops::matmul2d_descriptor::mode::multiply_accumulate),
-    mpp::tensor_ops::matmul2d<desc, execution_simdgroups<4>> mm;
+        execution_simdgroups<4>> mm;
 
     auto cT = mm.get_destination_cooperative_tensor<decltype(tA), decltype(tB), float>();
 #endif
@@ -8522,31 +8522,63 @@ kernel void kernel_mul_mm_id(
         ushort tiitg[[thread_index_in_threadgroup]],
         ushort tiisg[[thread_index_in_simdgroup]],
         ushort sgitg[[simdgroup_index_in_threadgroup]]) {
-
     threadgroup S0 * sa = (threadgroup S0 *)(shmem);
     threadgroup S1 * sb = (threadgroup S1 *)(shmem + 4096);
 
-    const int r0 = tgpig.y;
+    threadgroup float * sc = (threadgroup float *)(shmem);
-    const int r1 = tgpig.x;
+
+    constexpr int NR0 = 64;
+    constexpr int NR1 = 32;
+
+    constexpr int NK  = 32;
+    constexpr int NL0 = NK/16;
+    constexpr int NL1 = NK/8;
+
     const int im = tgpig.z; // expert
+    const int r0 = tgpig.y*NR0;
+    const int r1 = tgpig.x*NR1;
 
     device const uint32_t * tpe_u32 = (device const uint32_t *) (htpe);
     device const int32_t  * ids_i32 = (device const int32_t  *) (hids);
 
     const int32_t neh1 = tpe_u32[im];
 
-    if (r1*BLOCK_SIZE_N >= neh1) {
+    if (r1 >= neh1) {
         return;
     }
 
     // if this block is of 64x32 shape or smaller
-    const short n_rows = (args.ne0 - r0*BLOCK_SIZE_M < BLOCK_SIZE_M) ? (args.ne0 - r0*BLOCK_SIZE_M) : BLOCK_SIZE_M;
+    const short nr0 = (args.ne0 - r0 < NR0) ? (args.ne0 - r0) : NR0;
-    const short n_cols = (    neh1 - r1*BLOCK_SIZE_N < BLOCK_SIZE_N) ? (    neh1 - r1*BLOCK_SIZE_N) : BLOCK_SIZE_N;
+    const short nr1 = (    neh1 - r1 < NR1) ? (    neh1 - r1) : NR1;
 
     // a thread shouldn't load data outside of the matrix
-    const short thread_row = ((short)tiitg/THREAD_PER_ROW) < n_rows ? ((short)tiitg/THREAD_PER_ROW) : n_rows - 1;
+    const short lr0 = ((short)tiitg/NL0) < nr0 ? ((short)tiitg/NL0) : nr0 - 1; // 0 .. 63
-    const short thread_col = ((short)tiitg/THREAD_PER_COL) < n_cols ? ((short)tiitg/THREAD_PER_COL) : n_cols - 1;
+    const short lr1 = ((short)tiitg/NL1) < nr1 ? ((short)tiitg/NL1) : nr1 - 1; // 0 .. 31
 
+    const short il0 = (tiitg % NL0);
+
+    short il = il0;
+
+    const int id = ids_i32[im*args.ne21 + r1 + lr1];
+
+    const short i11 = (id % args.ne20) % args.ne11;
+    const short i12 = (id / args.ne20);
+    const short i13 = 0;
+
+    const uint64_t offset0 = im*args.nb02 + i13*args.nb03;
+    const short    offset1 = il0/nl;
+
+    device const block_q * x = (device const block_q *)(src0 + args.nb01*(r0 + lr0) + offset0) + offset1;
+
+    const short iy = 8*(tiitg % NL1);
+
+    device const T1 * y = (device const T1 *)(src1
+        + args.nb13*i13
+        + args.nb12*i12
+        + args.nb11*i11
+        + args.nb10*iy);
+
+#ifndef GGML_METAL_HAS_TENSOR
     S0_8x8 ma[4];
     S1_8x8 mb[2];
 
@@ -8555,39 +8587,36 @@ kernel void kernel_mul_mm_id(
     for (short i = 0; i < 8; i++){
         mc[i] = make_filled_simdgroup_matrix<float, 8>(0.f);
     }
+#else
+    auto tA = tensor<threadgroup S0, dextents<int32_t, 2>, tensor_inline>(sa, dextents<int32_t, 2>(NK,  NR0));
+    auto tB = tensor<threadgroup S1, dextents<int32_t, 2>, tensor_inline>(sb, dextents<int32_t, 2>(NR1, NK ));
 
-    short il = (tiitg % THREAD_PER_ROW);
+    mpp::tensor_ops::matmul2d<
+        mpp::tensor_ops::matmul2d_descriptor(NR1, NR0, NK, false, true, false, mpp::tensor_ops::matmul2d_descriptor::mode::multiply_accumulate),
+        execution_simdgroups<4>> mm;
 
-    const int id = ids_i32[im*args.ne21 + r1*BLOCK_SIZE_N + thread_col];
+    auto cT = mm.get_destination_cooperative_tensor<decltype(tA), decltype(tB), float>();
+#endif
 
-    const short i11 = (id % args.ne20) % args.ne11;
+    for (int loop_k = 0; loop_k < args.ne00; loop_k += NK) {
-    const short i12 = (id / args.ne20);
+#ifndef GGML_METAL_HAS_TENSOR
-    const short i13 = 0;
-
-    const uint64_t offset0 = im*args.nb02 + i13*args.nb03;
-    const short    offset1 = il/nl;
-
-    device const block_q * x = (device const block_q *)(src0
-        + args.nb01*(r0*BLOCK_SIZE_M + thread_row) + offset0) + offset1;
-
-    const short iy = (BLOCK_SIZE_K / THREAD_PER_COL * (tiitg % THREAD_PER_COL));
-
-    device const T1 * y = (device const T1 *)(src1
-        + args.nb13*i13
-        + args.nb12*i12
-        + args.nb11*i11
-        + args.nb10*iy);
-
-    for (int loop_k = 0; loop_k < args.ne00; loop_k += BLOCK_SIZE_K) {
         // load data and store to threadgroup memory
         if (is_same<T0_4x4, block_q>::value && FC_mul_mm_bc_inp) {
             threadgroup_barrier(mem_flags::mem_threadgroup);
 
             // no need for dequantization
             for (short i = 0; i < 16; i++) {
-                *(sa + SG_MAT_SIZE * ((tiitg/THREAD_PER_ROW/8) \
+                const short sx = 2*il0 + i/8;
-                +                     (tiitg%THREAD_PER_ROW)*16 + (i/8)*8) \
+                const short sy = (tiitg/NL0)/8;
-                +                     (tiitg/THREAD_PER_ROW)%8  + (i&7)*8) = loop_k + 16*il + i < args.ne00 ? ((device T0 *) x)[i] : 0;
+
+              //const short lx = i%8;
+              //const short ly = (tiitg/NL0)%8;
+                const short lx = (tiitg/NL0)%8;
+                const short ly = i%8;
+
+                const short ib = 8*sx + sy;
+
+                *(sa + 64*ib + 8*ly + lx) = loop_k + 16*il + i < args.ne00 ? *((device T0 *) x + i) : 0;
             }
         } else {
             S0_4x4 temp_a;
@@ -8596,85 +8625,188 @@ kernel void kernel_mul_mm_id(
             threadgroup_barrier(mem_flags::mem_threadgroup);
 
             FOR_UNROLL (short i = 0; i < 16; i++) {
-                *(sa + SG_MAT_SIZE * ((tiitg/THREAD_PER_ROW/8) \
+                const short sx = 2*il0 + i/8;
-                +                     (tiitg%THREAD_PER_ROW)*16 + (i/8)*8) \
+                const short sy = (tiitg/NL0)/8;
-                +                     (tiitg/THREAD_PER_ROW)%8  + (i&7)*8) = temp_a[i/4][i%4];
+
+              //const short lx = i%8;
+              //const short ly = (tiitg/NL0)%8;
+                const short lx = (tiitg/NL0)%8;
+                const short ly = i%8;
+
+                const short ib = 8*sx + sy;
+
+                // NOTE: this is massively slower.. WTF?
+                //sa[64*ib + 8*ly + lx] = temp_a[i/4][i%4];
+
+                *(sa + 64*ib + 8*ly + lx) = temp_a[i/4][i%4];
             }
         }
 
         if (FC_mul_mm_bc_inp) {
             for (short i = 0; i < 8; ++i) {
-                sb[32*8*(tiitg%THREAD_PER_COL) + 8*(tiitg/THREAD_PER_COL) + i] = loop_k + iy + i < args.ne00 ? (S1) ((device T1 *) y)[i] : 0;
+                const short sx = (tiitg%NL1);
+                const short sy = (tiitg/NL1)/8;
+
+                const short lx = i;
+                const short ly = (tiitg/NL1)%8;
+              //const short lx = (tiitg/NL1)%8;
+              //const short ly = i;
+
+                const short ib = 4*sx + sy;
+
+                *(sb + 64*ib + 8*ly + lx) = loop_k + iy + i < args.ne00 ? (S1) *((device T1 *) y + i) : 0;
             }
         } else {
-            *(threadgroup S1_2x4 *)(sb + 32*8*(tiitg%THREAD_PER_COL) + 8*(tiitg/THREAD_PER_COL)) = (S1_2x4)(*((device T1_2x4 *) y));
+            const short sx = (tiitg%NL1);
-        }
+            const short sy = (tiitg/NL1)/8;
 
-        il = (il + 2 < nl) ? il + 2 : il % 2;
+            const short dx = sx;
-        x  = (il < 2) ? x + (2 + nl - 1)/nl : x;
+            const short dy = sy;
-        y += BLOCK_SIZE_K;
+
+            const short ly = (tiitg/NL1)%8;
+
+            const short ib = 4*sx + sy;
+
+            *(threadgroup S1_2x4 *)(sb + 64*ib + 8*ly) = (S1_2x4)(*((device T1_2x4 *) y));
+        }
+#else
+        // load data and store to threadgroup memory
+        if (is_same<T0_4x4, block_q>::value && FC_mul_mm_bc_inp) {
+            threadgroup_barrier(mem_flags::mem_threadgroup);
+
+            // no need for dequantization
+            for (short i = 0; i < 16; i++) {
+                const short sx = 2*il0 + i/8;
+                const short sy = (tiitg/NL0)/8;
+
+                const short lx = i%8;
+                const short ly = (tiitg/NL0)%8;
+                //const short lx = (tiitg/NL0)%8;
+                //const short ly = i%8;
+
+                *(sa + NK*(8*sy + ly) + 8*sx + lx) = loop_k + 16*il + i < args.ne00 ? *((device T0 *) x + i) : 0;
+            }
+        } else {
+            S0_4x4 temp_a;
+            dequantize_func(x, il, temp_a);
 
             threadgroup_barrier(mem_flags::mem_threadgroup);
 
-        // load matrices from threadgroup memory and conduct outer products
+            FOR_UNROLL (short i = 0; i < 16; i++) {
-        threadgroup const S0 * lsma = (sa + THREAD_MAT_M*SG_MAT_SIZE*(sgitg%2));
+                const short sx = 2*il0 + i/8;
-        threadgroup const S1 * lsmb = (sb + THREAD_MAT_N*SG_MAT_SIZE*(sgitg/2));
+                const short sy = (tiitg/NL0)/8;
 
-        #pragma unroll(4)
+                const short lx = i%8;
-        for (short ik = 0; ik < BLOCK_SIZE_K/8; ik++) {
+                const short ly = (tiitg/NL0)%8;
-            #pragma unroll(4)
+                //const short lx = (tiitg/NL0)%8;
-            for (short i = 0; i < 4; i++) {
+                //const short ly = i%8;
-                simdgroup_load(ma[i], lsma + SG_MAT_SIZE * i);
+
+                *(sa + NK*(8*sy + ly) + 8*sx + lx) = temp_a[i/4][i%4];
+            }
+        }
+
+        if (FC_mul_mm_bc_inp) {
+            for (short i = 0; i < 8; ++i) {
+                const short sx = (tiitg%NL1);
+                const short sy = (tiitg/NL1)/8;
+
+                const short lx = i;
+                const short ly = (tiitg/NL1)%8;
+                //const short lx = (tiitg/NL1)%8;
+                //const short ly = i;
+
+                *(sb + NK*(8*sy + ly) + 8*sx + lx) = loop_k + iy + i < args.ne00 ? (S1) *((device T1 *) y + i) : 0;
+            }
+        } else {
+            const short sx = (tiitg%NL1);
+            const short sy = (tiitg/NL1)/8;
+
+            //const short lx = i;
+            const short ly = (tiitg/NL1)%8;
+            //const short lx = (tiitg/NL1)%8;
+            //const short ly = i;
+
+            *(threadgroup S1_2x4 *)(sb + NK*(8*sy + ly) + 8*sx) = (S1_2x4)(*((device T1_2x4 *) y));
+        }
+#endif
+
+        il = (il + 2 < nl) ? il + 2 : il % 2;
+        x  = (il < 2) ? x + (2 + nl - 1)/nl : x;
+
+        y += NK;
+
+        threadgroup_barrier(mem_flags::mem_threadgroup);
+
+#ifndef GGML_METAL_HAS_TENSOR
+        // load matrices from threadgroup memory and conduct outer products
+        threadgroup const S0 * lsma = (sa + 4*64*(sgitg%2));
+        threadgroup const S1 * lsmb = (sb + 2*64*(sgitg/2));
+
+        FOR_UNROLL (short ik = 0; ik < NK/8; ik++) {
+            simdgroup_barrier(mem_flags::mem_none);
+
+            FOR_UNROLL (short i = 0; i < 4; i++) {
+                simdgroup_load(ma[i], lsma + 64*i, 8, 0, false);
             }
 
             simdgroup_barrier(mem_flags::mem_none);
 
-            #pragma unroll(2)
+            FOR_UNROLL (short i = 0; i < 2; i++) {
-            for (short i = 0; i < 2; i++) {
+                simdgroup_load(mb[i], lsmb + 64*i, 8, 0, false);
-                simdgroup_load(mb[i], lsmb + SG_MAT_SIZE * i);
             }
 
-            #pragma unroll(8)
+            simdgroup_barrier(mem_flags::mem_none);
-            for (short i = 0; i < 8; i++){
+
+            FOR_UNROLL (short i = 0; i < 8; i++){
                 simdgroup_multiply_accumulate(mc[i], mb[i/4], ma[i%4], mc[i]);
             }
 
-            lsma += (BLOCK_SIZE_M/SG_MAT_ROW)*SG_MAT_SIZE;
+            lsma += 8*64;
-            lsmb += (BLOCK_SIZE_N/SG_MAT_ROW)*SG_MAT_SIZE;
+            lsmb += 4*64;
         }
+#else
+        auto sA = tA.slice(0, 0);
+        auto sB = tB.slice(0, 0);
+
+        mm.run(sB, sA, cT);
+#endif
     }
 
+    // block is smaller than 64x32, we should avoid writing data outside of the matrix
     threadgroup_barrier(mem_flags::mem_threadgroup);
 
-    threadgroup float * temp_str = ((threadgroup float *) shmem) \
+#ifdef GGML_METAL_HAS_TENSOR
-                                 + 32*(sgitg&1) + (16*(sgitg >> 1))*BLOCK_SIZE_M;
+    auto tC = tensor<threadgroup float, dextents<int32_t, 2>, tensor_inline>(sc, dextents<int32_t, 2>(NR0, NR1));
+    cT.store(tC);
+#else
+    threadgroup float * temp_str = ((threadgroup float *) shmem) + 32*(sgitg&1) + (16*(sgitg >> 1))*NR0;
 
-    #pragma unroll(8)
     for (short i = 0; i < 8; i++) {
-        simdgroup_store(mc[i], temp_str + 8*(i%4) + 8*BLOCK_SIZE_M*(i/4), BLOCK_SIZE_M);
+        simdgroup_store(mc[i], temp_str + 8*(i%4) + 8*NR0*(i/4), NR0, 0, false);
     }
+#endif
 
     threadgroup_barrier(mem_flags::mem_threadgroup);
 
-    for (short j = sgitg; j < n_cols; j += 4) {
+    for (short j = sgitg; j < nr1; j += 4) {
-        const int id = ids_i32[im*args.ne21 + r1*BLOCK_SIZE_N + j];
+        const int id = ids_i32[im*args.ne21 + r1 + j];
 
         const short ide = id % args.ne20;
         const short idt = id / args.ne20;
 
-        device float  * D  = (device float  *) dst + (r0*BLOCK_SIZE_M) + ide*args.ne0 + idt*args.ne1*args.ne0;
+        device float  * D  = (device float  *) dst + r0 + ide*args.ne0 + idt*args.ne1*args.ne0;
         device float4 * D4 = (device float4 *) D;
 
-        threadgroup float  * C  = (threadgroup float  *) shmem + (j*BLOCK_SIZE_M);
+        threadgroup float  * C  = (threadgroup float  *) shmem + j*NR0;
         threadgroup float4 * C4 = (threadgroup float4 *) C;
 
         int i = tiisg;
-        for (; i < n_rows/4; i += 32) {
+        for (; i < nr0/4; i += 32) {
             *(D4 + i) = *(C4 + i);
         }
 
-        i = (4*(n_rows/4)) + tiisg;
+        i = (4*(nr0/4)) + tiisg;
-        for (; i < n_rows; i += 32) {
+        for (; i < nr0; i += 32) {
             *(D + i) = *(C + i);
         }
     }