some numbers (third try)

Date view Thread view Subject view Author view

From: Godmar Back (gback@cs.utah.edu)
Date: Fri Dec 04 1998 - 15:28:50 EST

Forwarded message:
>From gback@cs.utah.edu Fri Dec 4 11:48:59 1998
X-UIDL: e2da5c3f8aba3cc4f486dc8184ebb40a
From: Godmar Back <gback@cs.utah.edu>
Message-Id: <199812041849.LAA11498@sal.cs.utah.edu>
Subject: some number (second try)
To: kaffe-core@rufus.w3.org (Kaffe Core Team)
Date: Fri, 4 Dec 1998 11:49:31 -0700 (MST)
Cc: Edouard.Parmelan@France.NCR.COM (Edouard Parmelan),
        gback@cs.utah.edu (Godmar Back)
X-Mailer: ELM [version 2.4 PL25]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

For the stackoverflow check.

Again, standard benchmark is compiling the run-time libraries,
I'm looking at the user time given by time to get a rough estimate.
 
JIT, no stackoverflow check:
      ca. 21 seconds.
  
JIT, stackoverflow check with one softcall and indirect call to
      threading interface:
      ca. 25 seconds.
  
JIT, stackoverflow check with one softcall and inlined jthread_on_current_stack:
      ca. 23.5 seconds.
  
>From these totally unrepresentative and unscientific numbers I conclude

a) stack overflow checking through a softcall introduces a significant slowdown.

b) inlining frequently used calls to the threading subsystem pays off.

    - Godmar

Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:05 EDT