Remove extraneous $ in inline asm for Intel syntax

[chirizxc]

Codspeed fails to build on nightly Rust with cranelift backend.

See rust-lang/rustc_codegen_cranelift#1601

[bjorn3]

The 32bit x86 variant also uses $$. cg_clif doesn't support 32bit x86 though.

[art049]

looks good thanks!

[codspeed-hq]

CodSpeed Performance Report

Merging #138 will degrade performances by 68.42%

_{Comparing chirizxc:fix-cranelift (8df58d4) with main (8923641)}

Summary

⚡ 16 improvements
❌ 22 regressions
✅ 325 untouched

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Mode	Benchmark	BASE	HEAD	Change
❌	WallTime	large_setup	18 ns	19 ns	-5.26%
⚡	Instrumentation	sleep_100ms	595.3 ns	566.1 ns	+5.15%
⚡	Instrumentation	sleep_10ms	595.3 ns	566.1 ns	+5.15%
⚡	Instrumentation	sleep_1ms	595.3 ns	566.1 ns	+5.15%
⚡	Instrumentation	sleep_50ms	595.3 ns	566.1 ns	+5.15%
⚡	WallTime	from_elem[2048]	264 ns	243 ns	+8.64%
⚡	WallTime	from_elem[4096]	371 ns	300 ns	+23.67%
⚡	WallTime	from_elem[8192]	550 ns	516 ns	+6.59%
❌	Instrumentation	bench_array1[10]	195.8 ns	225 ns	-12.96%
❌	Instrumentation	bench_array2[10]	195.8 ns	225 ns	-12.96%
❌	Instrumentation	mut_borrow	924.2 ns	953.3 ns	-3.06%
❌	WallTime	fib_10	375 ns	388 ns	-3.35%
❌	Instrumentation	recursive[0]	211.1 ns	240.3 ns	-12.14%
❌	Instrumentation	recursive[5]	291.1 ns	320.3 ns	-9.11%
❌	WallTime	recursive_memoized[BTreeMap<u64, u64>, 0]	14 ns	18 ns	-22.22%
⚡	Instrumentation	recursive_memoized[BTreeMap<u64, u64>, 0]	400.8 ns	371.7 ns	+7.85%
❌	WallTime	generate_combinations[5]	1.1 µs	1.1 µs	-7.86%
⚡	WallTime	graph_coloring[3]	2 µs	1.7 µs	+20.3%
❌	WallTime	graph_coloring[5]	1.1 µs	1.1 µs	-4.08%
❌	WallTime	hamiltonian_cycle[5]	863 ns	956 ns	-9.73%
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.