ggml-webgpu: address quantization precision and backend lifecycle managment (#21521)

* ggml(webgpu): fix the busy-polls in Emscripten  in the waitAny after #20618, and remove the busy webgpu log

* Merge with upstream

* Fix GET_ROWS packed integer NaN when using f16 as memory buffer in shader quants

* Update Unary wgsl EXP and EXPM1 for f16 stability

* Fix GET_ROWS IQ4_XS strcut for NaN f16 canonicalization

* Fix numerical percision for unary sqrt when working with f16

* Fix NaN canonicalization for packed integers using f16

* Update err threshold for binary div ops when using f16

* backend: Keep one Dawn/WebGPU instance alive for the lifetime of the static backend

* clean: uncomment existing code logs

* clean: clean the unncessary debug info

* Refactor and generalize dequant helpers

* Remove deprecated quant structs

* Refactor shader defines to reduce repetition

* Remove error override for F16 type

* fix: fix the accidential removal of the proper initialization of ctx

* clean: clean legacy and format code

* fix: did not modify tests ops

---------

Co-authored-by: Jeremy J. Hartmann <jeremy@mtion.tv>

ggml/src/ggml-webgpu/wgsl-shaders/unary.wgsl

@@ -107,7 +107,8 @@ fn main(@builtin(global_invocation_id) gid: vec3<u32>) {
     let res = src[params.offset_src + src_idx] / (1.0 + exp(-src[params.offset_src + src_idx]));
 #endif
 #ifdef EXP
-    let res = exp(src[params.offset_src + src_idx]);
+    let src_f32 = f32(src[params.offset_src + src_idx]);
+    let res = TYPE(exp(src_f32));
 #endif
 #ifdef LOG
     let res = TYPE(log(f32(src[params.offset_src + src_idx])));
@@ -161,7 +162,8 @@ fn main(@builtin(global_invocation_id) gid: vec3<u32>) {
     let res = TYPE(select(log(1.0 + exp(src_f32)), src_f32, src_f32 > 20.0));
 #endif
 #ifdef EXPM1
-    let res = exp(src[params.offset_src + src_idx]) - 1.0;
+    let src_f32 = f32(src[params.offset_src + src_idx]);
+    let res = TYPE(exp(src_f32) - 1.0);
 #endif
 #ifdef FLOOR
     let res = floor(src[params.offset_src + src_idx]);
@@ -181,7 +183,7 @@ fn main(@builtin(global_invocation_id) gid: vec3<u32>) {
     let res = src[params.offset_src + src_idx] * src[params.offset_src + src_idx];
 #endif
 #ifdef SQRT
-    let res = sqrt(src[params.offset_src + src_idx]);
+    let res = TYPE(sqrt(f32(src[params.offset_src + src_idx])));
 #endif
 #ifdef SIN
     let res_f32 = sin(f32(src[params.offset_src + src_idx]));