diff --git a/examples/quantize/quantize.cpp b/examples/quantize/quantize.cpp
index 694b70cf..51a66a8d 100644
--- a/examples/quantize/quantize.cpp
+++ b/examples/quantize/quantize.cpp
@@ -151,12 +151,14 @@ static bool try_parse_ftype(const std::string & ftype_str_in, llama_ftype & ftyp
 //
 [[noreturn]]
 static void usage(const char * executable) {
-    printf("usage: %s [--help] [--allow-requantize] [--leave-output-tensor] [--pure] [--imatrix] [--hide-imatrix] [--include-weights] [--exclude-weights] [--output-tensor-type] [--token-embedding-type] [--ffn-gate-inp-type] [--attn-q-type] [--attn-k-type] [--attn-v-type] [--attn-qkv-type] [--attn-output-type] [--ffn-gate-type] [--ffn-down-type] [--ffn-up-type] [--keep-split] [--partial-requant] [--override-kv] model-f32.gguf [model-quant.gguf] type [nthreads]\n\n", executable);
+    printf("usage: %s [--help] [--allow-requantize] [--leave-output-tensor] [--pure] [--imatrix] [--hide-imatrix] [--ignore-imatrix-rules] [--dry-run] [--include-weights] [--exclude-weights] [--output-tensor-type] [--token-embedding-type] [--ffn-gate-inp-type] [--attn-q-type] [--attn-k-type] [--attn-v-type] [--attn-qkv-type] [--attn-output-type] [--ffn-gate-type] [--ffn-down-type] [--ffn-up-type] [--repack] [--repack-pattern] [--keep-split] [--partial-requant] [--override-kv] model-f32.gguf [model-quant.gguf] type [nthreads]\n\n", executable);
     printf("  --allow-requantize: Allows requantizing tensors that have already been quantized. Warning: This can severely reduce quality compared to quantizing from 16bit or 32bit\n");
     printf("  --leave-output-tensor: Will leave output.weight un(re)quantized. Increases model size but may also increase quality, especially when requantizing\n");
     printf("  --pure: Disable k-quant mixtures and quantize all tensors to the same type\n");
     printf("  --imatrix file_name: use data in file_name as importance matrix for quant optimizations\n");
     printf("  --hide-imatrix: do not store imatrix details in the quantized model\n");
+    printf("  --ignore-imatrix-rules: ignore importance matrix rules when quantizing\n");
+    printf("  --dry-run: show what would be quantized without actually writing the output file\n");
     printf("  --include-weights tensor_name: use importance matrix for this/these tensor(s)\n");
     printf("  --exclude-weights tensor_name: use importance matrix for this/these tensor(s)\n");
     printf("  --output-tensor-type ggml_type: use this ggml_type for the output.weight tensor.\n");
diff --git a/ggml/src/iqk/iqk_flash_attn.cpp b/ggml/src/iqk/iqk_flash_attn.cpp
index 02509095..c0ece49d 100644
--- a/ggml/src/iqk/iqk_flash_attn.cpp
+++ b/ggml/src/iqk/iqk_flash_attn.cpp
@@ -169,15 +169,19 @@ extern "C" IQK_API bool iqk_flash_attn_noalibi(int type_q, int type_mask, float
 
     if (auto type_k = ggml_type(int_type_k_in), type_v = ggml_type(int_type_v); !are_kv_types_supported(type_k, type_v)) {
         if (ith == 0) {
-            fprintf(stderr, "\n==================== KV cache types %s, %s are not supported on the CPU\n",
+            fprintf(stderr, "\n==================== K cache %s coupled with V cache %s is not a supported combination on the CPU backend.\n",
                     ggml_type_name(type_k), ggml_type_name(type_v));
             auto & supported = supported_kv_types();
-            fprintf(stderr, "Sopprted types are:\n");
+            fprintf(stderr, "Supported types are:\n");
             for (auto type : supported) {
                 fprintf(stderr, "    %s\n", ggml_type_name(type));
             }
+            fprintf(stderr, "    Warning: ik_llama.cpp does not support Q5_0 or Q5_1 KV cache on the CPU.\n");
 #ifdef __AVX512BF16__
             fprintf(stderr, "    %s, but only if K and V are both %s\n", ggml_type_name(GGML_TYPE_BF16), ggml_type_name(GGML_TYPE_BF16));
+#endif
+#ifndef GGML_IQK_FA_ALL_QUANTS
+            fprintf(stderr, "    To enable q4_0, q4_1, and iq4_nl KV cache types, recompile with -DGGML_IQK_FA_ALL_QUANTS=ON\n");
 #endif
         }
         barrier(barrier_data);
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index 250e749f..3bf2e1ec 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -72,3 +72,7 @@ if (BUILD_SHARED_LIBS)
     set_target_properties(llama PROPERTIES POSITION_INDEPENDENT_CODE ON)
     target_compile_definitions(llama PRIVATE LLAMA_SHARED LLAMA_BUILD)
 endif()
+
+if (GGML_MAX_CONTEXTS)
+    add_compile_definitions(GGML_MAX_CONTEXTS=${GGML_MAX_CONTEXTS})
+endif()