Re: [libre-riscv-dev] Advanced Topics on RISCV
authorJean-Paul Chaput <Jean-Paul.Chaput@lip6.fr>
Tue, 24 Mar 2020 12:23:36 +0000 (13:23 +0100)
committerlibre-riscv-dev <libre-riscv-dev@lists.libre-riscv.org>
Tue, 24 Mar 2020 12:23:46 +0000 (12:23 +0000)
6f/43ac86b9a244302de1e8121b687f9c06311b9b [new file with mode: 0644]

diff --git a/6f/43ac86b9a244302de1e8121b687f9c06311b9b b/6f/43ac86b9a244302de1e8121b687f9c06311b9b
new file mode 100644 (file)
index 0000000..0f20cb7
--- /dev/null
@@ -0,0 +1,164 @@
+Return-path: <libre-riscv-dev-bounces@lists.libre-riscv.org>
+Envelope-to: publicinbox@libre-riscv.org
+Delivery-date: Tue, 24 Mar 2020 12:23:47 +0000
+Received: from localhost ([::1] helo=libre-riscv.org)
+       by libre-riscv.org with esmtp (Exim 4.89)
+       (envelope-from <libre-riscv-dev-bounces@lists.libre-riscv.org>)
+       id 1jGiaw-0002QH-KX; Tue, 24 Mar 2020 12:23:46 +0000
+Received: from isis.lip6.fr ([132.227.60.2])
+ by libre-riscv.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256)
+ (Exim 4.89) (envelope-from <Jean-Paul.Chaput@lip6.fr>)
+ id 1jGiav-0002QB-Pb
+ for libre-riscv-dev@lists.libre-riscv.org; Tue, 24 Mar 2020 12:23:45 +0000
+Received: from asim.lip6.fr (asim.lip6.fr [132.227.86.2])
+ by isis.lip6.fr (8.15.2/8.15.2) with ESMTPS id 02OCNjsu007285
+ (version=TLSv1.2 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO)
+ for <libre-riscv-dev@lists.libre-riscv.org>;
+ Tue, 24 Mar 2020 13:23:45 +0100 (CET)
+Received: from localhost (melon [132.227.103.13])
+ by asim.lip6.fr (8.15.2/8.14.4) with ESMTP id 02OCNib4011025
+ for <libre-riscv-dev@lists.libre-riscv.org>;
+ Tue, 24 Mar 2020 13:23:44 +0100 (MET)
+Message-ID: <686f9b58ae6f6a3e11caff9714db66a0406e8b0b.camel@lip6.fr>
+From: Jean-Paul Chaput <Jean-Paul.Chaput@lip6.fr>
+To: Libre-RISCV General Development <libre-riscv-dev@lists.libre-riscv.org>
+Date: Tue, 24 Mar 2020 13:23:36 +0100
+In-Reply-To: <0F80A1BC-0241-49FE-B374-B4F1D8F3F5BD@gatech.edu>
+References: <0F80A1BC-0241-49FE-B374-B4F1D8F3F5BD@gatech.edu>
+Organization: UPMC
+X-Mailer: Evolution 3.28.5 (3.28.5-5.el7) 
+Mime-Version: 1.0
+X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.4.3
+ (isis.lip6.fr [132.227.60.2]); Tue, 24 Mar 2020 13:23:45 +0100 (CET)
+X-Scanned-By: MIMEDefang 2.78 on 132.227.60.2
+X-Content-Filtered-By: Mailman/MimeDel 2.1.23
+Subject: Re: [libre-riscv-dev] Advanced Topics on RISCV
+X-BeenThere: libre-riscv-dev@lists.libre-riscv.org
+X-Mailman-Version: 2.1.23
+Precedence: list
+List-Id: Libre-RISCV General Development
+ <libre-riscv-dev.lists.libre-riscv.org>
+List-Unsubscribe: <http://lists.libre-riscv.org/mailman/options/libre-riscv-dev>, 
+ <mailto:libre-riscv-dev-request@lists.libre-riscv.org?subject=unsubscribe>
+List-Archive: <http://lists.libre-riscv.org/pipermail/libre-riscv-dev/>
+List-Post: <mailto:libre-riscv-dev@lists.libre-riscv.org>
+List-Help: <mailto:libre-riscv-dev-request@lists.libre-riscv.org?subject=help>
+List-Subscribe: <http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev>, 
+ <mailto:libre-riscv-dev-request@lists.libre-riscv.org?subject=subscribe>
+Reply-To: Libre-RISCV General Development
+ <libre-riscv-dev@lists.libre-riscv.org>
+Content-Type: multipart/mixed; boundary="===============1176392427724861224=="
+Errors-To: libre-riscv-dev-bounces@lists.libre-riscv.org
+Sender: "libre-riscv-dev" <libre-riscv-dev-bounces@lists.libre-riscv.org>
+
+
+--===============1176392427724861224==
+Content-Type: multipart/signed; micalg="pgp-sha1"; protocol="application/pgp-signature";
+       boundary="=-uMTCeTJLxmVTQeWr/JCW"
+
+
+--=-uMTCeTJLxmVTQeWr/JCW
+Content-Type: text/plain; charset="UTF-8"
+Content-Transfer-Encoding: quoted-printable
+
+
+
+Hello Immanuel,
+
+About manycores, we did a lot of work on that topic in our lab
+(Sorbonne Universit=C3=A9/LIP6). I send you a reference toward the
+project we developed. This is a huge lot of work and may take
+a lot of time to understand. But, we made a real chip with
+16 cores to prove it works (and it do work). I cannot give you
+more technical information as it not my field of expertise.
+
+You must distinguish two case:
+
+* Multi cores ( < 16) in that case, simpler solutions can be
+  used.
+
+* Many cores ( > 16) in this case, the TSAR project can be
+  of interest.
+
+TSAR (Tera-Scale ARchitecture).
+
+You can see details here:
+    https://www-soc.lip6.fr/trac/tsar
+
+Best regards,
+
+
+PS: Already signaled it to Luke.
+
+
+On Tue, 2020-03-24 at 11:51 +0000, Immanuel, Yehowshua U wrote:
+> I=E2=80=99ve read through the Spike page and a good portion of the simple=
+V page.
+>=20
+> My two goals at the moment are:
+> 1. Understand how RISCV handles multiple processes and does page walking
+> 2. Understand how multicore ROSCV would work
+>=20
+> I=E2=80=99m hoping to play with FreeRTOS soon so I can run through its co=
+debase for setting up
+> page tables.
+> Also, do you know if spike tests the special instructions like exception =
+instructions?
+> Also, what RISCV instructions would a kernel use to set up the pagetables=
+?
+>=20
+> Lastly, do you know any good resources for intro to multicore systems? RI=
+SCV doesn=E2=80=99t
+> seem to have any multicore specific instructions. My current questions wo=
+uld include
+> things like:
+>=20
+> 1. How can the kernel assign tasks to a certain core? If you have a proce=
+ss with
+> multiple threads, it would make sense to spread out the threads among ava=
+ilable
+> processors instead of concentrating them on a single core. How might this=
+ work with
+> respect to RISCV?
+>=20
+> 2. Does the hardware ensure cache coherency - that is - externally - soft=
+ware sees one
+> big cache all though I imagine each core would have a local cache that wo=
+uld have to
+> communicate with other caches?
+>=20
+> Yehowshua
+> _______________________________________________
+> libre-riscv-dev mailing list
+> libre-riscv-dev@lists.libre-riscv.org
+> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
+--=20
+
+      .-.     J e a n - P a u l   C h a p u t  /  Administrateur Systeme
+      /v\     Jean-Paul.Chaput@lip6.fr
+    /(___)\   work: (33) 01.44.27.53.99             =20
+     ^^ ^^    cell:      06.66.25.35.55   home: 09.65.29.83.38
+
+    U P M C   Universite Pierre & Marie Curie
+    L I P 6   Laboratoire d'Informatique de Paris VI
+    S o C     System On Chip
+
+--=-uMTCeTJLxmVTQeWr/JCW--
+
+
+
+--===============1176392427724861224==
+Content-Type: text/plain; charset="utf-8"
+MIME-Version: 1.0
+Content-Transfer-Encoding: base64
+Content-Disposition: inline
+
+X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KbGlicmUtcmlz
+Y3YtZGV2IG1haWxpbmcgbGlzdApsaWJyZS1yaXNjdi1kZXZAbGlzdHMubGlicmUtcmlzY3Yub3Jn
+Cmh0dHA6Ly9saXN0cy5saWJyZS1yaXNjdi5vcmcvbWFpbG1hbi9saXN0aW5mby9saWJyZS1yaXNj
+di1kZXYK
+
+--===============1176392427724861224==--
+
+
+