pack local_slot and constant in hir::Value
Instead of loading membase at the start of every function, just load it in HostToGuestThunk
vzeroupper in GuestToHostThunk before calling host function, and in HostToGuestThunk after calling function to prevent AVX dirty state slowdowns. In the future, check if CPU implements AVX as 128x2 and skip if so (https://john-h-k.github.io/VexTransitionPenalties.html)
Remove useless save/restore of ctx pointer, nothing modifies it and it prevents cpus from doing cross-function memory renaming (https://www.agner.org/forum/viewtopic.php?t=41). Could not remove the space on stack because of alignment issues, instead turned it into GUEST_SCRATCH64 which is a temporary that sequences may use
Reorder OpcodeInfo so that name is at offset 0, remove name and add GetOpcodeName function (name is only used for debug code, we are seperating frequently accessed data and rarely accessed data)
Add VECTOR_DENORMFLUSH opcode for handling output to DOT_PRODUCT and other opcodes that implicitly force denormal inputs/outputs to zero, will eventually use for implementing NJM
Rewrite sequences for LOAD_VECTOR_SHL/SHR. The mask with 0xf in it was pointless as all InstrEmit_ functions that create the load shift instructions do that in HIR. The tables are only used for nonzero constant inputs now, which are probably pretty rare. Instead of doing a shift and lookup, a base value is used for both in the constant table and adding/subtracting of the input is done
Reuse result of LoadVectorShl/Shr in InstrEmit_stvlx_, InstrEmit_stvrx_. We were previously calculating it twice which was contributing to the final sequences' fatness. Use OPCODE_SELECT instead of the sequence of or, andnot, and that it was using for merging
Add the proper unconditional denormal input flushing behavior to vfmadd, add it also to vfmsub (making the assumption it has the same behavior)
Remove constant propagation for DOT_PRODUCT_3/4
DOT_PRODUCT_3/4 now returns a vector with all four elements set to the result. (what we were doing before, truncating to float32 and then splatting didnt make any sense)
Add much more correct versions of DOT_PRODUCT_3/4, matching the Xb360's to 1 bit. Still needs work to be a perfect emulation.
Add constant folding for OPCODE_SELECT, OPCODE_INSERT, OPCODE_PERMUTE, OPCODE_SWIZZLE
Remove constant folding for DOT_PRODUCT
Removed the multibyte nop code I committed earlier, it doesnt help us much because nops are only used for debug stuff and its ugly and wouldnt survive in a pr to main
Check for AVX512BMI, use vpermb to shuffle if supported
Added recognition of impossible comparisons via NZM and optimize them away
Recognize (x + -y) and transform to (x - y) for constants
Recognize (~x ) + 1 and transform to -x
Check and transform comparisons if theyre semantically equal to others
Detect comparisons of single-bit values with their only possible non-zero value and transform to true/false tests
Transform ==0 to IS_FALSE, !=0 to IS_TRUE
Truncate to int8 if operand for IS_TRUE/IS_FALSE has a nzm of 1
Reduced code generated for SubDidCarry slightly
Add special case for InstrEmit_srawix if mask == 1
Cut down the code generated for trap instructions, instead of naive or'ing or compare results do a switch and select the best condition
Rerun simplification pass until no changes, as some optimizations will enable others to be done
Enable rel32 call optimization by default
Add special case to TYPE_INT64's EmitAnd for UINT_MAX mask. Do mov32 to 32 if detected to take advantage of implicit zero xt/reg renaming
Add helper function for skipping assignment defs in instr.
Add helper function for checking if an opcode is binary value type
Add several new optimizations to simplificationpass, plus weak NZM calculation code (better full evaluation of Z/NZ will be done later) .
List of optimizations:
If a value is anded with a bitmask that it was already masked against, reuse the old value (this cuts out most FPSCR update garbage, although it does cause a local variable to be allocated for the masked FPSCR and it still repeatedly stores the masked value to the context)
If masking a value that was or'ed against another check whether our mask only considers bits from one value or another. if so, change the operand to the OR input that actually matters
If the only usage of a rotate left's output is an AND against a mask that discards the bits that were rotated in change the opcode to SHIFT_LEFT
If masking against all ones, become an assign.
If XOR or OR against 0, become an assign (additional FPSCR codegen cleanup)
If XOR against all ones, become a NOT
Adding a direct CPUID check to x64_emitter for lzcnt, the version of xbyak we are using is skipping checking for lzcnt on all non-intel cpus, meaning we are generating the much slower bitscan path for AMD cpus.
the emitted asm as rel32 calls. Disabled by default, enabled via
resolve_rel32_guest_calls
detect whether cpu has fast jrcxz, fast loop/loope/loopne
much more thorough LoadConstantXMM
New cvar elide_e0_check that allows the backend to assume accesses via
the SP or TLS register will not cross into 0xe0 range
Add x64 codegen for Vector shift uint8
If has fast jrcxz use for some traptrue/breaktrue instructions
Use phat nops
Add cvar use_fast_dot_product, which uses a four instruction sequence
for both dot product instructions which ought to be equivalent. disabled
by default.
Replace a movzx after setae in both ComputeMemoryAddressOffset and ComputeMemoryAddress with a xor_ of eax prior to the cmp. This reduces the length in bytes of both sequences by 1, and should be a moderate ICache usage reduction thanks to the frequency of these sequences.
`pxor` is a zero-uop register-rename and `gf2p8affineqb dest, zero, int8`
is a very quick single-instruction way to use affine galois
transformations to fill a register with an immediate byte without
touching memory.
Edit one of the lanes in this unit-test to be larger than the width of
the element-size to ensure that this case is handled correctly.
It should only mask the lower `log2(32)=5` bits of the input, causing
`33`(`100001`) to be `1`(`000001`).
`vprolvd` is an almost 1:1 analog with this opcode and can be
conditionally emitted when the host supports AVX512{F,VL}.
Altivec docs say that `vrl{bhw}` masks the lower log2(n) bits of the
element-size.
[vprold](https://www.felixcloutier.com/x86/vprold:vprolvd:vprolq:vprolvq)
modulos the shift-value by the element size in bits, which is the same
as masking the lower log2(n) bits. So `vrlw` maps exactly to `vprold`.
[AltiVec](https://www.nxp.com/docs/en/reference-manual/ALTIVECPEM.pdf)
doc says that it just uses the lower `log2(n)` bits of the shift-amount
rather than the whole element-sized value. So there is no need to handle
an overflow. Also adjusts 64-bit literals to utilize the explicit
`UINT64_C` type.
In the `Int8` case of `VECTOR_SH{R,L}_V128`, when all the values are the
same, then a single-instruction `gf2p8affineqb` can be emitted that does
an int8-based arithmetic-shift, utilizing GF(8) arithmetic.
More info here:
https://wunkolo.github.io/post/2020/11/gf2p8affineqb-int8-shifting/
Also fixes the iteration-type for when detecting if all of the simd
lanes are the same value(was iterating `u16` and not `u8`)
Rather than having a huge list of if-statements that all do the same
thing, this preprocessor allows a more concise pattern to detecting if
the emit-flag is enabled as well as the correlated Xbyak flag that it
needs to check for to before allowing the feature-flag to be emitted.
Also moved the AVX-check to the beginning to early-out rather than do a
bunch of wasted work only to find out last that the host doesn't even
support AVX.
In the `Int8` case of `VECTOR_SHA_V128`, when all the values are the same, then a single-instruction `gf2p8affineqb` can be emitted that does an int8-based arithmetic-shift, utilizing GF(8) arithmetic.
More info here:
https://wunkolo.github.io/post/2020/11/gf2p8affineqb-int8-shifting/
As of now(Dec 2021): Tremont(Lakefield), Jasper Lake, Ice lake, Tigerlake, and Rocket Lake support GNFI.
Rather than having a single bool to conditionally detect haswell-level
instruction features. The granularity is increased with a new
`x64_extension_mask` where individual features within the x64 backend
can be turned on or off in a bit-mask manner. Since we have an ARM
backend on the horizon, I've added this to the new `x64`
configuration-group rather than `CPU`. This new pattern will hopefully
allow for testing to be more targetted to certain processor features and
allows the user to determine if they want certain features to be enabled
or disabled(such as avoiding BMI2 on certain AMD processors due to
pdep/pext being incredibly slow). The default configuration is to detect
and utilize all available features.
This addresses a JIT-issue in the case that the `src1` and `dest`
register are both the same. This issue only happens in the "generic"
x86 path but not in the BMI1-accelerated path.
Thanks Rick for the extensive debugging help.
When `src1` and `dest` were the same, then the `addc` instruction at
`82099A08` in title `584108FF` might emit the following assembly:
```
.text:82099A08 andc r11, r10, r11
|
| Jitted
|
V
00000000A0011B15 mov rbx,r10
00000000A0011B18 not rbx
00000000A0011B1B and rbx,rbx
```
This was due to the src1 operand and the destination register being the
same, which used to call the "else" case in the x64 emitter when it
needs to be handled explicitly due to register aliasing/allocation.
Addresses issue #1945
Fix the the case where src1 is constant and src2 is non-constant causing
an assert due to trying to call `.constant()` on the src2 operand.
Interfaces with an issue Gliniak was encountering where title `4D53082D`
encounters an assert. Also includes a BMI1-acceleration in the 64-bit
case where a temporary register is needed(the `and` x86 instruction only
supports immediate constants up to 32-bits).
In the case of having two register operands for `AndNot`, the `andn` instruction can be used when the host supports `BMI1`. `andn` only supports 32-bit and 64-bit operands, so some register up-casting is needed.
The `BMI1 feature` fits into the current pattern of `use_haswell_instructions` as BMI1 was only introduced in haswell.
Also moved the aliases to the end of the enum rather than interleave it with the bit definitions.
Implements the detection of some baseline `AVX512` subsets and some common aliases into `X64EmitterFeatureFlags`.
So far, `AVX512{F,VL,BW,DQ}` are the only subsets of `AVX512` that are detected with this PR since I anticipate these are the ones that will actually be used a lot in the x64 backend. Some aliases are also implemented such as `kX64EmitAVX512Ortho` which is `AVX512F` and `AVX512VL` combined which are the two subsets of AVX512 required to allow for `AVX512` operations upon `ymm` and `xmm` registers.
These aliases can possibly be collapsed since we could just always require `AVX512VL` to be supported to allow for _any_ kind of `AVX512` to be used since we will practically always want to use `AVX512` on `xmm` registers at the very least as there is no use-case where we want to use the 512-bit `zmm` registers exclusively.
Rather than using `n * 1024 * 1024`, this adds a convenient `_MiB`/`_KiB` user-literal to the new `literals.h` header to concisely describe units of memory in a much more readable way. Any other useful literals can be added to this header. These literals exist in the `xe::literals` namespace so they are opt-in, similar to `std::chrono` literals, and require a `using namespace xe::literals` statement to utilize it within the current scope.
I've done a pass through the codebase to replace trivial instances of `1024 * 1024 * ...` expressions being used but avoided anything that added additional casting complexity from `size_t` to `uint32_t` and such to keep this commit concise.
Just checking if the resulting mask is non-zero means we cannot allow this function to check for multiple features in parallel. A hypothetical computer that supports FMA but not AVX2 will return `true` if you try to call `IsFeatureEnabled(kX64EmitFMA | kX64EmitAVX2)`. We should make sure all the masked flags return `true` rather than check for non-zero.
This is ramping up to allow for particular subsets of AVX512 to be checked for in parallel with a single function call.
