search for: counters

...> > >> > >> > Summary >> > ======= >> > >> > 1. We should teach GlobalDCE to strip out instrumented functions which >> the >> > inliner cannot delete. >> > >> > 2. Sean suggests adding metadata to loads/stores of counters to convey >> that >> > they do not alias normal program data. >> > >> > I'm not familiar with AA, but off-hand this seems like a good idea. >> > >> > 3. Sean also suggests improving the optimizer to registerize counters >> when >...

Hi, I''m trying to execute something like this method in a model: def self.sql_for_rankings() sql = "SET @counter:=0;" sql << " SET @counter:=0; SELECT *, @counter:=@counter+1 AS rank FROM testscores. " find_by_sql(sql) end Where the new rank column is a kind of autoincrementer inside of

...l try to summarize the thread and respond > to some of the questions/feedback. > > > Summary > ======= > > 1. We should teach GlobalDCE to strip out instrumented functions which the > inliner cannot delete. > > 2. Sean suggests adding metadata to loads/stores of counters to convey that > they do not alias normal program data. > > I'm not familiar with AA, but off-hand this seems like a good idea. > > 3. Sean also suggests improving the optimizer to registerize counters when > possible. > > Ditto, seems like a good idea. &g...

There have been a lot of responses. I'll try to summarize the thread and respond to some of the questions/feedback. Summary ======= 1. We should teach GlobalDCE to strip out instrumented functions which the inliner cannot delete. 2. Sean suggests adding metadata to loads/stores of counters to convey that they do not alias normal program data. I'm not familiar with AA, but off-hand this seems like a good idea. 3. Sean also suggests improving the optimizer to registerize counters when possible. Ditto, seems like a good idea. 4. Sean has an alternate proposal for cou...

Hi All, I have a questions about associative list mappings in R, and if they are possible? I have data in the form show below, and want to make a new 'bucket' variable called combined. Which is the sum of the control and the exposed metric values This combined variable is a many to many matching as values only appear in the file if they have a value > 0. conversion.type filteredID

Hi, The current design of -fprofile-instr-generate has the same fundamental flaw as the old gcc's gcov instrumentation: it has contention on counters. A trivial synthetic test case was described here: http://lists.cs.uiuc.edu/pipermail/llvmdev/2013-October/066116.html For the problem to appear we need to have a hot function that is simultaneously executed by multiple threads -- then we will have high contention on the racy profile counters. Su...

I am acquiring some sampled data that is time-stamped (with a POSIXct). Some of the data is in the form of "counters" -- that is, what is interesting isn't value of a given counter at a given time, but the change in the counter from one sample to a later one. As the counters are only incremented, they would be perceived to be monotonically increasing -- ideally. Unfortunately, the counters can "wr...

On Thu, Apr 17, 2014 at 6:10 PM, Yaron Keren <yaron.keren at gmail.com> wrote: > If accuracy is not critical, incrementing the counters without any guards > might be good enough. > No. Contention on the counters leads to 5x-10x slowdown. This is never good enough. --kcc Hot areas will still be hot and cold areas will not be affected. > > Yaron > > > > 2014-04-17 15:21 GMT+03:00 Kostya Serebryany <kcc...

...ner runs and deletes f1, we're left with: void f2() { instrprof_increment(profc_f2); instrprof_increment(profc_f1); } Now we can say profc_f1 is redundant (or, an alias for profc_f2). I've noticed that frontend-based instrumentation can generate many redundant profile counters. This degrades performance and increases code size. We can address the problem by getting rid of redundant counter updates. The trick is to make sure we get back the same profiles. Proposed Solution ================= I propose a pruning pass which takes the following steps: 1. Delete functio...

...be happy to work with you on design issues and on code review. 3. Optimizing profile counter placement >From Eli's notes: > I remember we also spent some time discussing the counter intrinsics, and whether we could produce a different set of intrinsics in the frontend, and produce the counters later in the pipeline to avoid duplicate counters. I didn't completely follow that discussion; I haven't spent much time looking at the counter intrinsics or how they're lowered. Just to recap: the frontend emits calls to the llvm.instrprof_increment intrinsic to implement counter u...

Googling for this finds me a ton of stuff that doesn't actually do what I'm looking for. I want a simple hit counter cgi script of some kind that will increment a counter on every page load but I want to access the counter from a different page. In other words, I don't want to have a visible hit counter on the webpage itself. I want to have a different webpage that will show me the

...> 1. add some sort of alias annotation (such as an independent TBAA root >>> node) to all the counter increment memory instructions to tell the >>> optimizer they don't interfere with the usual loads and stores. >>> >>> 2. communicate to the optimizer that counters can be registerized. In a >>> loop like: >>> for (int i = 0; i < N; i++) { >>> if (foo()) >>> bar(); >>> else >>> baz(); >>> } >>> we perform O(N) counter increments (i.e. load, increment, store) last I >>&...

...of > my head, some are: > > 1. add some sort of alias annotation (such as an independent TBAA root > node) to all the counter increment memory instructions to tell the > optimizer they don't interfere with the usual loads and stores. > > 2. communicate to the optimizer that counters can be registerized. In a > loop like: > for (int i = 0; i < N; i++) { > if (foo()) > bar(); > else > baz(); > } > we perform O(N) counter increments (i.e. load, increment, store) last I > checked. However, if the counters are in registers, then we only per...

On Fri, Apr 18, 2014 at 12:13 AM, Dmitry Vyukov <dvyukov at google.com> wrote: > Hi, > > This is long thread, so I will combine several comments into single email. > > > >> - 8-bit per-thread counters, dumping into central counters on overflow. > >The overflow will happen very quickly with 8bit counter. > > Yes, but it reduces contention by 256x (a thread must execute at least 256 > loop iterations between increments). In practice, if you reduce contention > below some threshol...

...p, and I''ve hit a snag. Here''s a simplified example of the problem. (Excuse me if this example code isn''t strictly correct; I''m typing from memory.) Suppose I have a table which models a simple counter. It has a single column, "count": class CreateCounters < ActiveRecord::Migration def self.up create_table :counters do |t| t.column :count, :integer end end def self.down drop_table :counters end I''ve created a RESTful controller using the scaffold_resource generator. In the "show" view, I''d l...

Hi everybody! I've a large dataset, about 2 Mio entries of the format which I would like to import into a frame: <integer><integer><float><string><float><string><string> Because to the huge data amount I've choosen a binary format instead of a text format when exporting from Matlab. My import function is attached below. It works fine for only

I have an application where I need to send outbound prerecorded messages. The Asterisk "call file" process works fine if I am sending the call via SIP or IAX, but not via ZAP over a PRI channel. The destination device (my cell phone) never rings. The only unusual thing I see is on the fifth line of the "full" log, below, "channel.c: Don't know what to do with

https://bugzilla.netfilter.org/show_bug.cgi?id=1267 Bug ID: 1267 Summary: unable to use limit with counter in meters Product: nftables Version: unspecified Hardware: x86_64 OS: Gentoo Status: NEW Severity: enhancement Priority: P5 Component: nft Assignee: pablo at

Hi all, Does ENUMLOOKUP can query multiple DNS servers without having to replicate the same code in which the only thing replaced is the server? If I use ENUMLOOKUP(${exten}), Asterisk will parse enum.conf file to find the list of DNS servers in order of preference to be queried, but, I pretend to use something like this: ${ENUMLOOKUP(+${ARG1:2:},sip,c) which does not seam to care about

...>> On Fri, Apr 18, 2014 at 11:13 AM, Dmitry Vyukov <dvyukov at google.com>wrote: >> >>> Hi, >>> >>> This is long thread, so I will combine several comments into single >>> email. >>> >>> >>> >> - 8-bit per-thread counters, dumping into central counters on >>> overflow. >>> >The overflow will happen very quickly with 8bit counter. >>> >>> Yes, but it reduces contention by 256x (a thread must execute at least >>> 256 loop iterations between increments). In practice, if y...