From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Mon, 5 Jun 2006 05:17:09 +0400 From: Alexey Tourbin To: sisyphus@lists.altlinux.org Message-ID: <20060605011709.GC18120@localhost.localdomain> Mail-Followup-To: sisyphus@lists.altlinux.org References: <20060604210546.GA18120@localhost.localdomain> <200606041804.54270.iadzhubey@rics.bwh.harvard.edu> <20060604222630.GB18120@localhost.localdomain> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="jousvV0MzM2p6OtC" Content-Disposition: inline In-Reply-To: <20060604222630.GB18120@localhost.localdomain> Subject: Re: [sisyphus] ATLAS vs BLAS performance X-BeenThere: sisyphus@lists.altlinux.org X-Mailman-Version: 2.1.7 Precedence: list Reply-To: ALT Linux Sisyphus discussion list List-Id: ALT Linux Sisyphus discussion list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 05 Jun 2006 01:17:16 -0000 Archived-At: List-Archive: List-Post: --jousvV0MzM2p6OtC Content-Type: text/plain; charset=koi8-r Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Mon, Jun 05, 2006 at 02:26:30AM +0400, Alexey Tourbin wrote: > On Sun, Jun 04, 2006 at 06:04:54PM -0400, Ivan Adzhubey wrote: > > On Sunday 04 June 2006 17:05, Alexey Tourbin wrote: > > > =F1 =D2=C1=DA=CF=C2=D2=C1=CC=D3=D1, =CB=C1=CB =CE=C1=C4=CF =D3=CF=C2= =C9=D2=C1=D4=D8 ATLAS. > > > =F7=CF=D4 =D0=D2=C5=C4=D7=C1=D2=C9=D4=C5=CC=D8=CE=D9=CA benchmark. > > > > > > Fortran BLAS: > > > > mm <- matrix(rnorm(10^6), ncol =3D 10^3) > > > > system.time(crossprod(mm)) > > > > > > [1] 1.572 0.004 1.580 0.000 0.000 > > > > > > ATLAS w/ SSE2: > > > > mm <- matrix(rnorm(10^6), ncol =3D 10^3) > > > > system.time(crossprod(mm)) > > > > > > [1] 0.344 0.020 0.369 0.000 0.000 > >=20 > > =E1 =D7=CF=D4 =D3 GotoBLAS 1.2 (P4 2.8GHz, R 2.3.0): > >=20 > > > mm <- matrix(rnorm(10^6), ncol =3D 10^3) > > > system.time(crossprod(mm)) > > [1] 0.232 0.012 0.270 0.000 0.000 >=20 > model name : AMD Athlon(tm) 64 Processor 3200+ > cpu MHz : 2050.186 > cache size : 512 KB > =F7 =CF=C2=DD=C5=CD, =C5=D3=CC=C9 =DC=D4=CF =C2=D5=C4=C5=D4 =D2=C1=C2=CF= =D4=C1=D4=D8 =C9=DA =CB=CF=D2=CF=C2=CB=C9 =CB=C1=CB =CE=C1=C4=CF =C9 =C5=D3= =CC=C9 =D0=CF =D3=D2=C1=D7=CE=C5=CE=C9=C0 > =D3 GotoBLAS =D2=C1=DA=CE=C9=C3=C1 =C2=D5=C4=C5=D4 =CE=C5 =C8=D5=D6=C5, = =DE=C5=CD =D7 =D0=CF=CC=D4=CF=D2=C1 =D2=C1=DA=C1, =D4=CF=C7=C4=C1 =D1 =C2= =D5=C4=D5 > =C4=CF=D7=CF=CC=C5=CC=C5=CE. =F0=CF=D4=CF=CD=D5 =DE=D4=CF =CF=D3=CE=CF= =D7=CE=CF=CA =D2=C1=DA=D2=D9=D7 =D0=CF =D3=D2=C1=D7=CE=C5=CE=C9=C0 =D3 =C6= =CF=D2=D4=D2=C1=CE=CF=CD =D7=D3=A3 > =D2=C1=D7=CE=CF =CC=C9=CB=D7=C9=C4=C9=D2=CF=D7=C1=CE, =C1 10-20% =CE=C1 = =D0=D2=C1=CB=D4=C9=CB=C5 =D2=C5=C4=CB=CF =DE=D4=CF =D2=C5=DB=C1=C0=D4. =E1 =D7=CF=D4 ATLAS =C2=C5=DA =CF=D0=D4=C9=CD=C9=DA=C1=C3=C9=C9 =D0=CF=C4 P= 4SSE2: > mm <- matrix(rnorm(10^6), ncol =3D 10^3) > system.time(crossprod(mm)) [1] 0.584 0.012 0.624 0.000 0.000 =F0=CF=DE=D4=C9 =D7 =C4=D7=C1 =D2=C1=DA=C1 =C8=D5=D6=C5, =DE=C5=CD =D3 SSE2= (=CE=CF =D0=CF=DE=D4=C9 =D7 =D4=D2=C9 =D2=C1=DA=C1 =CC=D5=DE=DB=C5, =DE=C5= =CD =C6=CF=D2=D4=D2=C1=CE). =FA=CE=C1=DE=C9=D4 =D3=D4=CF=C9=D4 =C4=C5=CC=C1=D4= =D8 /usr/lib/sse2. --jousvV0MzM2p6OtC Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (GNU/Linux) iD8DBQFEg4YVfBKgtDjnu0YRAktiAKDPblYFx/jFZm3Cqg0rw+FVoVbnzgCgisrh zSRhvFI3T1tmsG3sObA4Ysc= =PiK2 -----END PGP SIGNATURE----- --jousvV0MzM2p6OtC--