Instruction
HRESET
Resets the parts of the processor's prediction history that EAX selects and the operating system enabled in IA32_HRESET_ENABLE; allowed only in ring 0.
- Extensions
- Forms
- 1 (1 Legacy)
- Data
- Intel XED v2026.08.23
HRESET asks the processor to forget history it keeps about the code running on this logical processor. Intel’s manuals define one kind of it so far: the history behind Intel Thread Director, which watches the running software thread and reports a class for it in IA32_THREAD_FEEDBACK_CHAR, so that the operating system can place the thread on a core type that suits it. A kernel can reset it when it switches threads, so that what the previous thread did does not color the class of the next.
EAX picks, the immediate does not
The components to reset are the set bits of EAX; bit 0 is Thread Director’s history, and the rest are reserved. The 8-bit immediate is ignored. EAX = 0 makes HRESET a NOP. Three levels have to agree:
- CPUID leaf 07H, subleaf 1, EAX bit 22 says that HRESET exists, or it raises #UD;
- CPUID leaf 20H, EBX, lists the components this processor can reset;
- the kernel enables some of those in the IA32_HRESET_ENABLE MSR, which takes no others, and a bit of EAX that the MSR leaves clear raises #GP(0).
HRESET runs only at CPL 0, or raises #GP(0), and it aborts a TSX
transaction. With -mhreset, or an -march such as alderlake, arrowlake or
pantherlake (not sapphirerapids), GCC 14 and Clang 23 at -O2 compile
_hreset(x) to a move into EAX and hreset 0.
History that goes anyway
Intel’s manual lists the times the processor resets Thread Director’s history by itself: on entry to and exit from SMM when IA32_DEBUGCTL.FREEZE_WHILE_SMM is set, at GETSEC[SENTER] (on every logical processor) and GETSEC[ENTERACCS], and at INIT or the wait for SIPI. An HRESET also clears the valid bit of IA32_THREAD_FEEDBACK_CHAR, which tells the kernel to ignore the class until the hardware has gathered enough to report one again.
The encoding
HRESET is F3 0F 3A F0 with the ModRM byte fixed at C0, then the immediate. With a VEX prefix, the same map and opcode is RORX.
AMD’s manual defines no HRESET, and the Zen 5 guest that ran the probe did not report it in CPUID, so this page has no recorded cases.
Forms
| Instruction | Encoding | Requires |
|---|---|---|
HRESET imm8
HRESET_IMM8 | F3 0F 3A F0 C0 ib | HRESET |
Latency and throughput
| Model | imm8 |
|---|---|
| Alder Lake | 100 / 0.25p0156 |
Operation
What the forms do, in pseudocode that our tests run against results recorded on real processors. How to read.
if CPL > 0:
fault(GP, 0)
if eax & ~msr[0x17DA]:
fault(GP, 0) # a bit that IA32_HRESET_ENABLE
# leaves clear
pass # forgets the history of each
# component that EAX selects: bit 0,
# Thread Director's; imm8 is ignored
Not run on a processor: the Zen 5 guest that ran the probe does not report HRESET in CPUID.
Operands
| Operand | Kind | Action |
|---|---|---|
HRESET imm8 | ||
imm8 | explicit | read |
EAX | implicit | read |
Flags
None: XED records no flag that the forms test or change.
Attributes
What else XED records about the forms. How to read.
- All forms
- privileged: CPL 0 only
Intrinsics
The C intrinsics of Clang 23.1.2 that compile to HRESET, each with the form it compiles to and the Clang options it needs. The forms come from compiling every intrinsic and decoding the result with Intel XED; all intrinsics A–Z explains the method.
| Intrinsic | Compiles to |
|---|---|
_hreset
void _hreset(int __eax) |
|
CPUID and processors
Each form belongs to an XED ISA set. A form can be used when the processor reports every CPUID bit of one of its ISA set's alternatives.
| ISA set | CPUID | Processors |
|---|---|---|
HRESET | HRESET | Alder Lake, Arrow Lake, Panther Lake, Nova Lake, Granite Rapids, Diamond Rapids, Sierra Forest, Clearwater Forest |
Sources
- Intel XED v2026.08.23 (commit
0bcb6237345c): forms, encodings, operands, flags, exception classes, attributes, ISA sets, CPUID bits and chip model. - Clang's x86 intrinsic headers, LLVM 23.1.2 (commit
85ac56026243): the intrinsics, their signatures and doc comments, and the Clang of that release, which compiled each of them. - LLVM 23.1.2's x86 scheduling models (commit
85ac56026243): the latency, reciprocal throughput and ports of each form, as llvm-mca reports them. - Intel 64 and IA-32 Architectures Software Developer's Manual, Volume 2 (325383-093, September 2026): Volume 2A, page 3-440 (HRESET)
- Intel 64 and IA-32 Architectures Software Developer's Manual, Volume 1 (253665-093, September 2026)
- Intel 64 and IA-32 Architectures Software Developer's Manual, Volume 3 (325384-093, September 2026)
- Intel Architecture Instruction Set Extensions and Future Features Programming Reference (319433-062, June 2026)
The tables are derived from Intel XED, Copyright Intel Corporation, licensed under the Apache License 2.0; x86-64.net converted and reformatted them. The intrinsics and their signatures come from Clang and the timings from LLVM's scheduling models, both part of the LLVM Project, licensed under the Apache License 2.0 with LLVM Exceptions. The text is our own.