@nockdev/awf 6.2.6 → 6.2.7

package/.agent/skills/languages/asm/references/advanced-architectures.md

@@ -0,0 +1,191 @@
+## 🏛️ Legacy Assemblers & DOS
+
+### Assembler Comparison
+
+| Assembler | Syntax    | Linker | 16-bit | Direct .COM | Best For              |
+| --------- | --------- | ------ | ------ | ----------- | --------------------- |
+| **NASM**  | Intel     | LD/GCC | ✅    | ✅         | Cross-platform        |
+| **MASM**  | Intel     | LINK   | ✅    | ❌          | Windows/Visual Studio |
+| **GAS**   | AT&T      | LD     | ✅    | ❌          | GCC/Linux ecosystem   |
+| **TASM**  | Intel     | TLINK  | ✅    | ✅         | DOS + Turbo Debugger  |
+| **FASM**  | Intel     | None   | ✅    | ✅         | Self-contained, fast  |
+| **A86**   | Custom    | None   | ✅    | ✅         | Simple DOS programs   |
+| **YASM**  | NASM-like | LD     | ✅    | ✅         | NASM alternative      |
+| **HLA**   | High-lvl  | Link   | ❌     | ❌          | Learning ASM          |
+
+### 16-bit Real Mode (8086/8088)
+
+```asm
+; NASM 16-bit DOS .COM program
+org 0x100              ; .COM files load at 100h
+
+section .text
+start:
+    mov ah, 09h        ; DOS print string
+    mov dx, message
+    int 21h
+
+    mov ah, 4Ch        ; DOS exit
+    int 21h
+
+section .data
+message db "Hello, DOS!", 13, 10, "$"
+```
+
+### DOSBox Development Setup
+
+```yaml
+setup:
+  1. Install DOSBox/DOSBox-X
+  2. Mount work directory: mount c c:\asm
+  3. Install assembler (TASM/MASM/NASM)
+  4. Use DEBUG.EXE or Turbo Debugger
+
+debuggers:
+  DEBUG.EXE: "Built-in DOS debugger (line-oriented)"
+  Turbo Debugger: "Full-screen, source-level debugging"
+  D86: "A86's companion debugger"
+```
+
+### CPU Evolution (16 → 32 → 64-bit)
+
+| CPU     | Year | Bits | Address | Mode         | Memory |
+| ------- | ---- | ---- | ------- | ------------ | ------ |
+| 8086    | 1978 | 16   | 20-bit  | Real         | 1 MB   |
+| 8088    | 1979 | 16   | 20-bit  | Real         | 1 MB   |
+| 80286   | 1982 | 16   | 24-bit  | Real/Prot    | 16 MB  |
+| 80386   | 1985 | 32   | 32-bit  | Real/Prot/V8 | 4 GB   |
+| 80486   | 1989 | 32   | 32-bit  | Real/Prot/V8 | 4 GB   |
+| Pentium | 1993 | 32   | 32-bit  | Real/Prot/V8 | 4 GB   |
+| x64     | 2003 | 64   | 48-bit  | Long         | 256 TB |
+
+---
+
+## 🆕 RISC-V Assembly
+
+### Register Map (RV64I)
+
+| ABI Name | Reg    | Role            | Preserved |
+| -------- | ------ | --------------- | --------- |
+| zero     | x0     | Hard-wired zero | N/A       |
+| ra       | x1     | Return address  | No        |
+| sp       | x2     | Stack pointer   | Yes       |
+| gp       | x3     | Global pointer  | N/A       |
+| tp       | x4     | Thread pointer  | N/A       |
+| t0-t2    | x5-7   | Temporaries     | No        |
+| s0/fp    | x8     | Saved/Frame ptr | Yes       |
+| s1       | x9     | Saved register  | Yes       |
+| a0-a7    | x10-17 | Args / Return   | No        |
+| s2-s11   | x18-27 | Saved registers | Yes       |
+| t3-t6    | x28-31 | Temporaries     | No        |
+
+### Calling Convention
+
+```asm
+# RISC-V function example
+my_function:
+    addi sp, sp, -16     # Allocate stack
+    sd ra, 8(sp)         # Save return address
+    sd s0, 0(sp)         # Save s0
+
+    # ... function body ...
+    # Args in a0-a7, return in a0/a1
+
+    ld s0, 0(sp)         # Restore s0
+    ld ra, 8(sp)         # Restore ra
+    addi sp, sp, 16      # Deallocate
+    ret
+```
+
+### Common Extensions
+
+| Ext | Description         |
+| --- | ------------------- |
+| M   | Multiply/Divide     |
+| A   | Atomics             |
+| F   | Single-precision FP |
+| D   | Double-precision FP |
+| C   | Compressed (16-bit) |
+| V   | Vector extension    |
+
+---
+
+## 🆕 Intel AVX10 & AMX (2025-2026)
+
+### AVX10.2 (June 2025)
+
+```yaml
+features:
+  - Mixed-precision: FP16, BF16 operations
+  - IEEE-754-2019 NaN propagation
+  - Unified P-core/E-core support
+  - Embedded rounding in 256-bit mode
+  - 32 vector registers (ZMM0-31)
+
+detection:
+  CPUID: Check AVX10 version via CPUID.07H
+```
+
+### Intel AMX (Sapphire Rapids+)
+
+```yaml
+description: "Accelerate matrix operations for AI/ML"
+
+tiles:
+  count: 8 (TMM0-TMM7)
+  max_size: 16 rows × 64 bytes
+
+data_types:
+  - BF16 (Brain Float 16)
+  - INT8 (8-bit integer)
+  - FP16 (Half precision)
+
+speedup: "Up to 10x for deep learning"
+```
+
+### AMX Intrinsics Example
+
+```c
+#include <immintrin.h>
+
+void amx_matmul_bf16(void* c, void* a, void* b) {
+    _tile_loadd(0, a, 64);   // Load A → TMM0
+    _tile_loadd(1, b, 64);   // Load B → TMM1
+    _tile_dpbf16ps(2, 0, 1); // C += A × B (BF16)
+    _tile_stored(2, c, 64);  // Store result
+}
+```
+
+---
+
+## 🛠️ Modern IDE Tools (2025-2026)
+
+### VS Code Extensions
+
+| Extension                 | Purpose                |
+| ------------------------- | ---------------------- |
+| `ms-vscode.cpptools`      | C/C++ IntelliSense     |
+| `maziac.asm-code-lens`    | ASM syntax + debugging |
+| `dan-c-underwood.arm`     | ARM assembly support   |
+| `13xforever.language-x86` | x86/x64 syntax         |
+| `Cortex-Debug`            | ARM embedded debugging |
+
+### ARM-Specific Tools
+
+| Tool           | Description                     |
+| -------------- | ------------------------------- |
+| ArmLS          | ARM64 language server (LSP)     |
+| Arm Dev Studio | Full IDE + Performance Analyzer |
+| QEMU           | ARM64/RISC-V emulation          |
+
+### Online Assembly Tools
+
+| Tool               | URL                      |
+| ------------------ | ------------------------ |
+| Compiler Explorer  | godbolt.org              |
+| ASM Editor (specy) | asm-editor.specy.app     |
+| onlinegdb          | onlinegdb.com/online_asm |
+
+---
+
+_DOMYH Awesome Code Assembly Language Skill Comprehensive Guide_

package/.agent/skills/languages/asm/references/build-structure.md

@@ -0,0 +1,150 @@
+## 📄 Program Structure Templates
+
+### NASM (Linux x64)
+
+```asm
+; program.asm
+section .data
+    msg db "Hello, World!", 10
+    len equ $ - msg
+
+section .bss
+    buffer resb 256
+
+section .text
+    global _start
+
+_start:
+    ; sys_write(1, msg, len)
+    mov rax, 1          ; syscall: write
+    mov rdi, 1          ; fd: stdout
+    mov rsi, msg        ; buffer
+    mov rdx, len        ; count
+    syscall
+
+    ; sys_exit(0)
+    mov rax, 60         ; syscall: exit
+    xor rdi, rdi        ; status: 0
+    syscall
+```
+
+### MASM (Windows x64)
+
+```asm
+; program.asm
+.data
+    msg db "Hello, World!", 0
+
+.code
+    externdef printf:proc
+
+main PROC
+    sub rsp, 40         ; Shadow space + alignment
+    lea rcx, msg        ; First arg
+    call printf
+    add rsp, 40
+    xor eax, eax        ; Return 0
+    ret
+main ENDP
+
+END
+```
+
+### GAS (ARM64)
+
+```asm
+// program.s
+.data
+msg:    .ascii "Hello, World!\n"
+len = . - msg
+
+.text
+.global _start
+
+_start:
+    mov x8, #64         // sys_write
+    mov x0, #1          // fd: stdout
+    ldr x1, =msg        // buffer
+    mov x2, #len        // count
+    svc #0
+
+    mov x8, #93         // sys_exit
+    mov x0, #0          // status
+    svc #0
+```
+
+---
+
+## 🔨 Build Commands
+
+### NASM + LD (Linux)
+
+```bash
+# Compile
+nasm -f elf64 program.asm -o program.o
+
+# Link
+ld program.o -o program
+
+# Or with C runtime
+nasm -f elf64 program.asm -o program.o
+gcc program.o -o program -no-pie
+```
+
+### MASM + MSVC (Windows)
+
+```powershell
+# Compile ASM
+ml64 /c program.asm /Fo program.obj
+
+# Link with C++ code
+cl /c main.cpp
+link main.obj program.obj /OUT:main.exe
+
+# Or combined in Visual Studio project
+# Just add .asm file after enabling MASM build customization
+```
+
+### GAS + GCC (Cross-platform)
+
+```bash
+# Compile
+as program.s -o program.o
+
+# Link
+ld program.o -o program
+
+# Or with GCC
+gcc -c program.s -o program.o
+gcc main.c program.o -o main
+```
+
+### CMake Integration
+
+```cmake
+# Enable ASM
+enable_language(ASM_NASM)
+# or
+enable_language(ASM_MASM)
+
+# Add ASM sources
+add_executable(myprogram
+    main.cpp
+    asm_routines.asm
+)
+```
+
+---
+
+## ✅ Best Practices Checklist
+
+- [ ] **Avoid inline ASM if possible** Use intrinsics, compiler optimizations
+- [ ] **Profile before optimizing** Don't guess, measure
+- [ ] **Keep ASM sections minimal** Maintain only critical paths
+- [ ] **Use runtime CPU detection** Support multiple instruction sets
+- [ ] **Align data properly** 16/32/64-byte for SIMD operations
+- [ ] **Follow calling conventions** Preserve callee-saved registers
+- [ ] **Document thoroughly** ASM is harder to understand
+- [ ] **Test on target hardware** SIMD behavior varies
+
+---

package/.agent/skills/languages/asm/references/simd-programming.md

@@ -0,0 +1,92 @@
+## ⚡ SIMD Programming
+
+### Intrinsics vs Inline ASM Decision
+
+```yaml
+recommendation:
+  use_intrinsics:
+    - Portable across compilers
+    - Compiler can optimize around them
+    - Easier to maintain
+    - Better for SSE/AVX/AVX-512
+
+  use_inline_asm:
+    - CPU-specific features not in intrinsics
+    - Precise instruction control needed
+    - Very small, critical code paths
+```
+
+### SSE/AVX Intrinsics Example
+
+```c
+#include <immintrin.h>
+
+// Add 8 floats using AVX
+void vector_add_avx(float* result, const float* a, const float* b) {
+    __m256 va = _mm256_load_ps(a);    // Load 8 floats from a
+    __m256 vb = _mm256_load_ps(b);    // Load 8 floats from b
+    __m256 vr = _mm256_add_ps(va, vb); // Add
+    _mm256_store_ps(result, vr);       // Store result
+}
+
+// Dot product using SSE
+float dot_product_sse(const float* a, const float* b, int n) {
+    __m128 sum = _mm_setzero_ps();
+    for (int i = 0; i < n; i += 4) {
+        __m128 va = _mm_load_ps(&a[i]);
+        __m128 vb = _mm_load_ps(&b[i]);
+        sum = _mm_add_ps(sum, _mm_mul_ps(va, vb));
+    }
+    // Horizontal add
+    sum = _mm_hadd_ps(sum, sum);
+    sum = _mm_hadd_ps(sum, sum);
+    return _mm_cvtss_f32(sum);
+}
+```
+
+### Runtime CPU Feature Detection
+
+```c
+#include <cpuid.h>  // GCC/Clang
+
+bool has_avx2(void) {
+    unsigned int eax, ebx, ecx, edx;
+    __cpuid(7, eax, ebx, ecx, edx);
+    return (ebx & (1 << 5)) != 0;  // AVX2 bit
+}
+
+// Dispatch based on CPU features
+void process_data(float* data, int n) {
+    if (has_avx2()) {
+        process_avx2(data, n);
+    } else if (has_sse4()) {
+        process_sse4(data, n);
+    } else {
+        process_scalar(data, n);
+    }
+}
+```
+
+### ARM64 NEON Intrinsics
+
+```c
+#include <arm_neon.h>
+
+// Add 4 floats using NEON
+void vector_add_neon(float* result, const float* a, const float* b) {
+    float32x4_t va = vld1q_f32(a);     // Load 4 floats
+    float32x4_t vb = vld1q_f32(b);     // Load 4 floats
+    float32x4_t vr = vaddq_f32(va, vb); // Add
+    vst1q_f32(result, vr);              // Store
+}
+
+// Horizontal sum
+float horizontal_sum_neon(float32x4_t v) {
+    float32x2_t sum = vadd_f32(vget_low_f32(v), vget_high_f32(v));
+    sum = vpadd_f32(sum, sum);
+    return vget_lane_f32(sum, 0);
+}
+```
+
+---
+

package/.agent/skills/languages/c/META.yaml

@@ -1,5 +1,5 @@
 name: c
-version: "6.2.6"
+version: "6.2.7"
 display: C
 category: language
 tier: 3