Return-Path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references; bh=qH+dPKh/W/jSl0rHt0xqNAbHjXfBz2C0nCif1Vm7yqU=; b=hFO5ioZQ3Yjd8GxjzJ9CS73wDeMJ82hEtZhRQLm12IY6wJ+S+bXHls+52rXFkYDRMj moqnJY9SXOVoJ6y3Tf2Q4rrYTaO0SNyBXiZ4o5QjP3h+qaB2cN1xt/fhzZVVEHlXdFZz eWJ2ye0dQeX/peFAYMBwLyKzXMtIaPWoiZJeCEIAbUY9jfw0lk3nCuRWwaqeJ/SOX1Y9 xer9GhMMLZf5bWZahHeus6qwAV4BdNuEcTsyV2GhhGuXKEuaSvyolC/f56VaXWgKjlhh EDg6xGUIdnw2SNewQ0dIXxvd4uaBykjVjes3/34w8iEnb5Ac2GTRMAN46IFAENDWezoU uuEQ== From: SeongJae Park Subject: [PATCH 3/3] memorder: Present P0()-P3() in consistent manner Date: Wed, 20 Sep 2017 00:34:56 +0900 Message-Id: <20170919153456.10274-3-sj38.park@gmail.com> In-Reply-To: <20170919153456.10274-1-sj38.park@gmail.com> References: <20170919153456.10274-1-sj38.park@gmail.com> To: paulmck@linux.vnet.ibm.com Cc: perfbook@vger.kernel.org, SeongJae Park List-ID: Signed-off-by: SeongJae Park --- memorder/memorder.tex | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/memorder/memorder.tex b/memorder/memorder.tex index 85ba46e..5df5649 100644 --- a/memorder/memorder.tex +++ b/memorder/memorder.tex @@ -1982,12 +1982,12 @@ This sequence of events will depend critically on \co{P0()} and Figure~\ref{fig:memorder:Shared Store Buffers And Multi-Copy Atomicity}. \QuickQuiz{} - But just how is it fair that P0 and P1 must share a store buffer - and a cache, but P2 gets one each of its very own??? + But just how is it fair that \co{P0()} and \co{P1()} must share a store + buffer and a cache, but \co{P2()} gets one each of its very own??? \QuickQuizAnswer{ - Presumably there is a P3, as is in fact shown in + Presumably there is a \co{P3()}, as is in fact shown in Figure~\ref{fig:memorder:Shared Store Buffers And Multi-Copy Atomicity}, - that shares P2's store buffer and cache. + that shares \co{P2()}'s store buffer and cache. But not necessarily. Some platforms allow different cores to disable different numbers of threads, allowing the hardware to adjust to the needs of the @@ -2008,8 +2008,8 @@ Figure~\ref{fig:memorder:Shared Store Buffers And Multi-Copy Atomicity}. } \QuickQuizEnd Row~1 shows the initial state, with the initial value of \co{y} in -P0's and P1's shared cache, and the initial value of \co{x} in -P2's cache. +\co{P0()}'s and \co{P1()}'s shared cache, and the initial value of \co{x} in +\co{P2()}'s cache. Row~2 shows the immediate effect of \co{P0()} executing its store on line~8. Because the cacheline containing \co{x} is not in \co{P0()}'s and \co{P1()}'s -- 2.10.0