[Qemu-devel] [PATCH 0/3] target/riscv: use tcg_lookup_and_goto_ptr

[Qemu-devel] [PATCH 0/3] target/riscv: use tcg_lookup_and_goto_ptr

[Emilio G. Cota]

There are a few more places where the lookup could be inserted,
but I think these are the ones that will matter performance-wise.

Perf results in patch 3's log.

Regarding the benchmarks: I'd have used SPEC06 but I don't have
much time to get it to compile. Is there a guide on how to do so?
Ideally I'd like to use the fedora image, which works very well.

Thanks,

		Emilio

[Qemu-devel] [PATCH 1/3] target/riscv: optimize cross-page direct jumps in softmmu

[Emilio G. Cota]

Signed-off-by: Emilio G. Cota <cota@braap.org>
---
 target/riscv/translate.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 0b6be74f2d..ec2988b4f6 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -135,7 +135,7 @@ static void gen_goto_tb(DisasContext *ctx, int n, target_ulong dest)
         if (ctx->base.singlestep_enabled) {
             gen_exception_debug();
         } else {
-            tcg_gen_exit_tb(NULL, 0);
+            tcg_gen_lookup_and_goto_ptr();
         }
     }
 }
-- 
2.17.1

Re: [Qemu-devel] [PATCH 1/3] target/riscv: optimize cross-page direct jumps in softmmu

[Richard Henderson]

On 08/09/2018 02:43 PM, Emilio G. Cota wrote:
> Signed-off-by: Emilio G. Cota <cota@braap.org>
> ---
>  target/riscv/translate.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 0b6be74f2d..ec2988b4f6 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -135,7 +135,7 @@ static void gen_goto_tb(DisasContext *ctx, int n, target_ulong dest)
>          if (ctx->base.singlestep_enabled) {
>              gen_exception_debug();
>          } else {
> -            tcg_gen_exit_tb(NULL, 0);
> +            tcg_gen_lookup_and_goto_ptr();
>          }
>      }
>  }
> 

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>


r~

[Qemu-devel] [PATCH 2/3] target/riscv: optimize indirect branches

[Emilio G. Cota]

Signed-off-by: Emilio G. Cota <cota@braap.org>
---
 target/riscv/translate.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index ec2988b4f6..66a80ca772 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -548,7 +548,7 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
         if (rd != 0) {
             tcg_gen_movi_tl(cpu_gpr[rd], ctx->pc_succ_insn);
         }
-        tcg_gen_exit_tb(NULL, 0);
+        tcg_gen_lookup_and_goto_ptr();
 
         if (misaligned) {
             gen_set_label(misaligned);
-- 
2.17.1

Re: [Qemu-devel] [PATCH 2/3] target/riscv: optimize indirect branches

[Richard Henderson]

On 08/09/2018 02:43 PM, Emilio G. Cota wrote:
> Signed-off-by: Emilio G. Cota <cota@braap.org>
> ---
>  target/riscv/translate.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index ec2988b4f6..66a80ca772 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -548,7 +548,7 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
>          if (rd != 0) {
>              tcg_gen_movi_tl(cpu_gpr[rd], ctx->pc_succ_insn);
>          }
> -        tcg_gen_exit_tb(NULL, 0);
> +        tcg_gen_lookup_and_goto_ptr();
>  
>          if (misaligned) {
>              gen_set_label(misaligned);
> 


Reviewed-by: Richard Henderson <richard.henderson@linaro.org>


r~

[Qemu-devel] [PATCH 3/3] target/riscv: call tcg_lookup_and_goto_ptr on DISAS_TOO_MANY

[Emilio G. Cota]

Performance impact of this and the previous commit, measured with
the very-easy-to-cross-compile rv8-bench:
  https://github.com/rv8-io/rv8-bench

Host: Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz

- Key:
  before: master
  after1,2,3: the 3 commits in this series (i.e. 3 is this commit)

- User-mode:

 bench      before  after1  after2  after3  final speedup
---------------------------------------------------------
 aes        1.12s   1.12s   1.10s   1.00s   1.12
 bigint     0.78s   0.78s   0.78s   0.78s   1
 dhrystone  0.96s   0.97s   0.49s   0.49s   1.9591837
 miniz      1.94s   1.94s   1.88s   1.86s   1.0430108
 norx       0.51s   0.51s   0.49s   0.48s   1.0625
 primes     0.85s   0.85s   0.84s   0.84s   1.0119048
 qsort      4.87s   4.88s   1.86s   1.86s   2.6182796
 sha512     0.76s   0.77s   0.64s   0.64s   1.1875

(after1 only applies to softmmu, so no surprises here)

- Full-system (fedora):

 bench      before  after1  after2  after3  final speedup
---------------------------------------------------------
 aes        2.68s   2.54s   2.60s   2.34s   1.1452991
 bigint     1.61s   1.56s   1.55s   1.64s   0.98170732
 dhrystone  1.78s   1.67s   1.25s   1.24s   1.4354839
 miniz      3.53s   3.35s   3.28s   3.35s   1.0537313
 norx       1.13s   1.09s   1.07s   1.06s   1.0660377
 primes     15.37s  15.41s  15.20s  15.37s  1
 qsort      7.20s   6.71s   3.85s   3.96s   1.8181818
 sha512     1.07s   1.04s   0.90s   0.90s   1.1888889

SoftMMU slows things down, so the numbers are less sensitive.
Cross-page jumps improve things a little bit, though.

Note that I'm not showing here averages, just results from a
single run, so with primes there isn't much to worry about.

Signed-off-by: Emilio G. Cota <cota@braap.org>
---
 target/riscv/translate.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 66a80ca772..98e0311606 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1872,7 +1872,7 @@ static void riscv_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu)
         if (ctx->base.singlestep_enabled) {
             gen_exception_debug();
         } else {
-            tcg_gen_exit_tb(NULL, 0);
+            tcg_gen_lookup_and_goto_ptr();
         }
         break;
     case DISAS_NORETURN:
-- 
2.17.1

Re: [Qemu-devel] [PATCH 3/3] target/riscv: call tcg_lookup_and_goto_ptr on DISAS_TOO_MANY

[Richard Henderson]

On 08/09/2018 02:43 PM, Emilio G. Cota wrote:
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 66a80ca772..98e0311606 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1872,7 +1872,7 @@ static void riscv_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu)
>          if (ctx->base.singlestep_enabled) {
>              gen_exception_debug();
>          } else {
> -            tcg_gen_exit_tb(NULL, 0);
> +            tcg_gen_lookup_and_goto_ptr();
>          }
>          break;
>      case DISAS_NORETURN:

This could just as easily use

  case DISAS_TOO_MANY:
    gen_goto_tb(ctx, 0, ctx->base.pc_next);


r~