From: Alistair Francis <alistair.francis@opensource.wdc.com>
To: qemu-devel@nongnu.org
Cc: alistair23@gmail.com, Frank Chang <frank.chang@sifive.com>,
Richard Henderson <richard.henderson@linaro.org>,
Alistair Francis <alistair.francis@wdc.com>
Subject: [PULL 34/88] target/riscv: rvv-1.0: take fractional LMUL into vector max elements calculation
Date: Mon, 20 Dec 2021 14:56:11 +1000 [thread overview]
Message-ID: <20211220045705.62174-35-alistair.francis@opensource.wdc.com> (raw)
In-Reply-To: <20211220045705.62174-1-alistair.francis@opensource.wdc.com>
From: Frank Chang <frank.chang@sifive.com>
Update vext_get_vlmax() and MAXSZ() to take fractional LMUL into
calculation for RVV 1.0.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20211210075704.23951-27-frank.chang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
---
target/riscv/cpu.h | 27 ++++++++++++++++---------
target/riscv/cpu_helper.c | 16 ++++++++++++---
target/riscv/insn_trans/trans_rvv.c.inc | 12 ++++++++++-
3 files changed, 42 insertions(+), 13 deletions(-)
diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index 709b7c3abb..11a0f41b27 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -430,18 +430,27 @@ static inline RISCVMXL riscv_cpu_mxl(CPURISCVState *env)
#endif
/*
- * A simplification for VLMAX
- * = (1 << LMUL) * VLEN / (8 * (1 << SEW))
- * = (VLEN << LMUL) / (8 << SEW)
- * = (VLEN << LMUL) >> (SEW + 3)
- * = VLEN >> (SEW + 3 - LMUL)
+ * Encode LMUL to lmul as follows:
+ * LMUL vlmul lmul
+ * 1 000 0
+ * 2 001 1
+ * 4 010 2
+ * 8 011 3
+ * - 100 -
+ * 1/8 101 -3
+ * 1/4 110 -2
+ * 1/2 111 -1
+ *
+ * then, we can calculate VLMAX = vlen >> (vsew + 3 - lmul)
+ * e.g. vlen = 256 bits, SEW = 16, LMUL = 1/8
+ * => VLMAX = vlen >> (1 + 3 - (-3))
+ * = 256 >> 7
+ * = 2
*/
static inline uint32_t vext_get_vlmax(RISCVCPU *cpu, target_ulong vtype)
{
- uint8_t sew, lmul;
-
- sew = FIELD_EX64(vtype, VTYPE, VSEW);
- lmul = FIELD_EX64(vtype, VTYPE, VLMUL);
+ uint8_t sew = FIELD_EX64(vtype, VTYPE, VSEW);
+ int8_t lmul = sextract32(FIELD_EX64(vtype, VTYPE, VLMUL), 0, 3);
return cpu->cfg.vlen >> (sew + 3 - lmul);
}
diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
index 1b31d0ad47..10f3baba53 100644
--- a/target/riscv/cpu_helper.c
+++ b/target/riscv/cpu_helper.c
@@ -75,12 +75,22 @@ void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,
*cs_base = 0;
if (riscv_has_ext(env, RVV)) {
+ /*
+ * If env->vl equals to VLMAX, we can use generic vector operation
+ * expanders (GVEC) to accerlate the vector operations.
+ * However, as LMUL could be a fractional number. The maximum
+ * vector size can be operated might be less than 8 bytes,
+ * which is not supported by GVEC. So we set vl_eq_vlmax flag to true
+ * only when maxsz >= 8 bytes.
+ */
uint32_t vlmax = vext_get_vlmax(env_archcpu(env), env->vtype);
- bool vl_eq_vlmax = (env->vstart == 0) && (vlmax == env->vl);
+ uint32_t sew = FIELD_EX64(env->vtype, VTYPE, VSEW);
+ uint32_t maxsz = vlmax << sew;
+ bool vl_eq_vlmax = (env->vstart == 0) && (vlmax == env->vl) &&
+ (maxsz >= 8);
flags = FIELD_DP32(flags, TB_FLAGS, VILL,
FIELD_EX64(env->vtype, VTYPE, VILL));
- flags = FIELD_DP32(flags, TB_FLAGS, SEW,
- FIELD_EX64(env->vtype, VTYPE, VSEW));
+ flags = FIELD_DP32(flags, TB_FLAGS, SEW, sew);
flags = FIELD_DP32(flags, TB_FLAGS, LMUL,
FIELD_EX64(env->vtype, VTYPE, VLMUL));
flags = FIELD_DP32(flags, TB_FLAGS, VL_EQ_VLMAX, vl_eq_vlmax);
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index e12db9aae8..5c04ac90da 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -1049,7 +1049,17 @@ GEN_LDST_WHOLE_TRANS(vs8r_v, 8, true)
/*
*** Vector Integer Arithmetic Instructions
*/
-#define MAXSZ(s) (s->vlen >> (3 - s->lmul))
+
+/*
+ * MAXSZ returns the maximum vector size can be operated in bytes,
+ * which is used in GVEC IR when vl_eq_vlmax flag is set to true
+ * to accerlate vector operation.
+ */
+static inline uint32_t MAXSZ(DisasContext *s)
+{
+ int scale = s->lmul - 3;
+ return scale < 0 ? s->vlen >> -scale : s->vlen << scale;
+}
static bool opivv_check(DisasContext *s, arg_rmrr *a)
{
--
2.31.1
next prev parent reply other threads:[~2021-12-20 5:53 UTC|newest]
Thread overview: 90+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-20 4:55 [PULL 00/88] riscv-to-apply queue Alistair Francis
2021-12-20 4:55 ` [PULL 01/88] target/riscv: zfh: half-precision load and store Alistair Francis
2021-12-20 4:55 ` [PULL 02/88] target/riscv: zfh: half-precision computational Alistair Francis
2021-12-20 4:55 ` [PULL 03/88] target/riscv: zfh: half-precision convert and move Alistair Francis
2021-12-20 4:55 ` [PULL 04/88] target/riscv: zfh: half-precision floating-point compare Alistair Francis
2021-12-20 4:55 ` [PULL 05/88] target/riscv: zfh: half-precision floating-point classify Alistair Francis
2021-12-20 4:55 ` [PULL 06/88] target/riscv: zfh: add Zfh cpu property Alistair Francis
2021-12-20 4:55 ` [PULL 07/88] target/riscv: zfh: implement zfhmin extension Alistair Francis
2021-12-20 4:55 ` [PULL 08/88] target/riscv: zfh: add Zfhmin cpu property Alistair Francis
2021-12-20 4:55 ` [PULL 09/88] target/riscv: drop vector 0.7.1 and add 1.0 support Alistair Francis
2021-12-20 4:55 ` [PULL 10/88] target/riscv: Use FIELD_EX32() to extract wd field Alistair Francis
2021-12-20 4:55 ` [PULL 11/88] target/riscv: rvv-1.0: add mstatus VS field Alistair Francis
2021-12-20 4:55 ` [PULL 12/88] target/riscv: rvv-1.0: set mstatus.SD bit if mstatus.VS is dirty Alistair Francis
2021-12-20 4:55 ` [PULL 13/88] target/riscv: rvv-1.0: add sstatus VS field Alistair Francis
2021-12-20 4:55 ` [PULL 14/88] target/riscv: rvv-1.0: introduce writable misa.v field Alistair Francis
2021-12-20 4:55 ` [PULL 15/88] target/riscv: rvv-1.0: add translation-time vector context status Alistair Francis
2021-12-20 4:55 ` [PULL 16/88] target/riscv: rvv-1.0: remove rvv related codes from fcsr registers Alistair Francis
2021-12-20 4:55 ` [PULL 17/88] target/riscv: rvv-1.0: add vcsr register Alistair Francis
2021-12-20 4:55 ` [PULL 18/88] target/riscv: rvv-1.0: add vlenb register Alistair Francis
2021-12-20 4:55 ` [PULL 19/88] target/riscv: rvv-1.0: check MSTATUS_VS when accessing vector csr registers Alistair Francis
2021-12-20 4:55 ` [PULL 20/88] target/riscv: rvv-1.0: remove MLEN calculations Alistair Francis
2021-12-20 4:55 ` [PULL 21/88] target/riscv: rvv-1.0: add fractional LMUL Alistair Francis
2021-12-20 4:55 ` [PULL 22/88] target/riscv: rvv-1.0: add VMA and VTA Alistair Francis
2021-12-20 4:56 ` [PULL 23/88] target/riscv: rvv-1.0: update check functions Alistair Francis
2021-12-20 4:56 ` [PULL 24/88] target/riscv: introduce more imm value modes in translator functions Alistair Francis
2021-12-20 4:56 ` [PULL 25/88] target/riscv: rvv:1.0: add translation-time nan-box helper function Alistair Francis
2021-12-20 4:56 ` [PULL 26/88] target/riscv: rvv-1.0: remove amo operations instructions Alistair Francis
2021-12-20 4:56 ` [PULL 27/88] target/riscv: rvv-1.0: configure instructions Alistair Francis
2021-12-20 4:56 ` [PULL 28/88] target/riscv: rvv-1.0: stride load and store instructions Alistair Francis
2021-12-20 4:56 ` [PULL 29/88] target/riscv: rvv-1.0: index " Alistair Francis
2021-12-20 4:56 ` [PULL 30/88] target/riscv: rvv-1.0: fix address index overflow bug of indexed load/store insns Alistair Francis
2021-12-20 4:56 ` [PULL 31/88] target/riscv: rvv-1.0: fault-only-first unit stride load Alistair Francis
2021-12-20 4:56 ` [PULL 32/88] target/riscv: rvv-1.0: load/store whole register instructions Alistair Francis
2021-12-20 4:56 ` [PULL 33/88] target/riscv: rvv-1.0: update vext_max_elems() for load/store insns Alistair Francis
2021-12-20 4:56 ` Alistair Francis [this message]
2021-12-20 4:56 ` [PULL 35/88] target/riscv: rvv-1.0: floating-point square-root instruction Alistair Francis
2021-12-20 4:56 ` [PULL 36/88] target/riscv: rvv-1.0: floating-point classify instructions Alistair Francis
2021-12-20 4:56 ` [PULL 37/88] target/riscv: rvv-1.0: count population in mask instruction Alistair Francis
2021-12-20 4:56 ` [PULL 38/88] target/riscv: rvv-1.0: find-first-set mask bit instruction Alistair Francis
2021-12-20 4:56 ` [PULL 39/88] target/riscv: rvv-1.0: set-X-first mask bit instructions Alistair Francis
2021-12-20 4:56 ` [PULL 40/88] target/riscv: rvv-1.0: iota instruction Alistair Francis
2021-12-20 4:56 ` [PULL 41/88] target/riscv: rvv-1.0: element index instruction Alistair Francis
2021-12-20 4:56 ` [PULL 42/88] target/riscv: rvv-1.0: allow load element with sign-extended Alistair Francis
2021-12-20 4:56 ` [PULL 43/88] target/riscv: rvv-1.0: register gather instructions Alistair Francis
2021-12-20 4:56 ` [PULL 44/88] target/riscv: rvv-1.0: integer scalar move instructions Alistair Francis
2021-12-20 4:56 ` [PULL 45/88] target/riscv: rvv-1.0: floating-point move instruction Alistair Francis
2021-12-20 4:56 ` [PULL 46/88] target/riscv: rvv-1.0: floating-point scalar move instructions Alistair Francis
2021-12-20 4:56 ` [PULL 47/88] target/riscv: rvv-1.0: whole register " Alistair Francis
2021-12-20 4:56 ` [PULL 48/88] target/riscv: rvv-1.0: integer extension instructions Alistair Francis
2021-12-20 4:56 ` [PULL 49/88] target/riscv: rvv-1.0: single-width averaging add and subtract instructions Alistair Francis
2021-12-20 4:56 ` [PULL 50/88] target/riscv: rvv-1.0: single-width bit shift instructions Alistair Francis
2021-12-20 4:56 ` [PULL 51/88] target/riscv: rvv-1.0: integer add-with-carry/subtract-with-borrow Alistair Francis
2021-12-20 4:56 ` [PULL 52/88] target/riscv: rvv-1.0: narrowing integer right shift instructions Alistair Francis
2021-12-20 4:56 ` [PULL 53/88] target/riscv: rvv-1.0: widening integer multiply-add instructions Alistair Francis
2021-12-20 4:56 ` [PULL 54/88] target/riscv: rvv-1.0: single-width saturating add and subtract instructions Alistair Francis
2021-12-20 4:56 ` [PULL 55/88] target/riscv: rvv-1.0: integer comparison instructions Alistair Francis
2021-12-20 4:56 ` [PULL 56/88] target/riscv: rvv-1.0: floating-point compare instructions Alistair Francis
2021-12-20 4:56 ` [PULL 57/88] target/riscv: rvv-1.0: mask-register logical instructions Alistair Francis
2021-12-20 4:56 ` [PULL 58/88] target/riscv: rvv-1.0: slide instructions Alistair Francis
2021-12-20 4:56 ` [PULL 59/88] target/riscv: rvv-1.0: floating-point " Alistair Francis
2021-12-20 4:56 ` [PULL 60/88] target/riscv: rvv-1.0: narrowing fixed-point clip instructions Alistair Francis
2021-12-20 4:56 ` [PULL 61/88] target/riscv: rvv-1.0: single-width floating-point reduction Alistair Francis
2021-12-20 4:56 ` [PULL 62/88] target/riscv: rvv-1.0: widening floating-point reduction instructions Alistair Francis
2021-12-20 4:56 ` [PULL 63/88] target/riscv: rvv-1.0: single-width scaling shift instructions Alistair Francis
2021-12-20 4:56 ` [PULL 64/88] target/riscv: rvv-1.0: remove widening saturating scaled multiply-add Alistair Francis
2021-12-20 4:56 ` [PULL 65/88] target/riscv: rvv-1.0: remove vmford.vv and vmford.vf Alistair Francis
2021-12-20 4:56 ` [PULL 66/88] target/riscv: rvv-1.0: remove integer extract instruction Alistair Francis
2021-12-20 4:56 ` [PULL 67/88] target/riscv: rvv-1.0: floating-point min/max instructions Alistair Francis
2021-12-20 4:56 ` [PULL 68/88] target/riscv: introduce floating-point rounding mode enum Alistair Francis
2021-12-20 4:56 ` [PULL 69/88] target/riscv: rvv-1.0: floating-point/integer type-convert instructions Alistair Francis
2021-12-20 4:56 ` [PULL 70/88] target/riscv: rvv-1.0: widening floating-point/integer type-convert Alistair Francis
2021-12-20 4:56 ` [PULL 71/88] target/riscv: add "set round to odd" rounding mode helper function Alistair Francis
2021-12-20 4:56 ` [PULL 72/88] target/riscv: rvv-1.0: narrowing floating-point/integer type-convert Alistair Francis
2021-12-20 4:56 ` [PULL 73/88] target/riscv: rvv-1.0: relax RV_VLEN_MAX to 1024-bits Alistair Francis
2021-12-20 4:56 ` [PULL 74/88] target/riscv: rvv-1.0: implement vstart CSR Alistair Francis
2021-12-20 4:56 ` [PULL 75/88] target/riscv: rvv-1.0: trigger illegal instruction exception if frm is not valid Alistair Francis
2021-12-20 4:56 ` [PULL 76/88] target/riscv: gdb: support vector registers for rv64 & rv32 Alistair Francis
2021-12-20 4:56 ` [PULL 77/88] target/riscv: rvv-1.0: floating-point reciprocal square-root estimate instruction Alistair Francis
2021-12-20 4:56 ` [PULL 78/88] target/riscv: rvv-1.0: floating-point reciprocal " Alistair Francis
2021-12-20 4:56 ` [PULL 79/88] target/riscv: rvv-1.0: rename r2_zimm to r2_zimm11 Alistair Francis
2021-12-20 4:56 ` [PULL 80/88] target/riscv: rvv-1.0: add vsetivli instruction Alistair Francis
2021-12-20 4:56 ` [PULL 81/88] target/riscv: rvv-1.0: add evl parameter to vext_ldst_us() Alistair Francis
2021-12-20 4:56 ` [PULL 82/88] target/riscv: rvv-1.0: add vector unit-stride mask load/store insns Alistair Francis
2021-12-20 4:57 ` [PULL 83/88] target/riscv: rvv-1.0: rename vmandnot.mm and vmornot.mm to vmandn.mm and vmorn.mm Alistair Francis
2021-12-20 4:57 ` [PULL 84/88] target/riscv: rvv-1.0: update opivv_vadc_check() comment Alistair Francis
2021-12-20 4:57 ` [PULL 85/88] target/riscv: rvv-1.0: Add ELEN checks for widening and narrowing instructions Alistair Francis
2021-12-20 4:57 ` [PULL 86/88] riscv: Set 5.4 as minimum kernel version for riscv32 Alistair Francis via
2021-12-20 4:57 ` [PULL 87/88] target/riscv: Enable bitmanip Zb[abcs] instructions Alistair Francis
2021-12-20 4:57 ` [PULL 88/88] hw/riscv: Use load address rather than entry point for fw_dynamic next_addr Alistair Francis
2021-12-20 21:19 ` [PULL 00/88] riscv-to-apply queue Richard Henderson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20211220045705.62174-35-alistair.francis@opensource.wdc.com \
--to=alistair.francis@opensource.wdc.com \
--cc=alistair.francis@wdc.com \
--cc=alistair23@gmail.com \
--cc=frank.chang@sifive.com \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).