How can I modify the cache architecture in the gem5 simulator? I want to be able to configure the cache so that it is used differently (either set-associative or fully associative) depending on a certain execution mode?
I am new to gem5 and still finding my way. Where can I change the modelling of the cache architecture and how it is used in the source files?
Thanks!
You can change the modelling of the cache architecture by changing the Tags. The Tags are instantiated in src/mem/cache/cache.py.
This modification can be done either by:
Creating a config file. It is the correct approach and will give you more control over the simulation (check Jason's tutorial http://learning.gem5.org/book/part1/simple_config.html to learn how to create a config file).
Quick and dirty approach by modifying the default value of Tags between BaseSetAssoc and FullyAssoc within the cache.py file. This, however, will not work towards your final goal of changing the model according to the execution mode.
Related
When using CSScript.Evaluator.Reset(), will this reset anyone else that is currently using the same script or build?
Another way of asking my question.
script = CSScript.Evaluator.LoadCode(scriptString);
If another user came thru and called the same code, but an error occurred.
Would the original compile still be good and safe?
Also CSScript.Evaluator.AutoResetEvaluatorOnError = true, this does not seem to be working.
After i cause an error in the code, then fix it. The compile will not work til i reset my app, or use Reset().
Using Reset is working, But that's the reason for my initial question.
I am using CSScript.Evaluator.LoadCode and looking at CSScript.Evaluator.LoadMethod, but getting the same issues.
These are not huge scripts, but may run in large batches or loops.
Not against unique naming, cause i will have build for every run anyway.
But i'm not sure the cache is working either.
Where is the location of the cache folder, when CSScript.CacheEnabled is enabled?
Would the original compile still be good and safe?
Yes it would as when you hold the reference to the compiled object it is YOURS. I is good even if you destroy the compiler.
Also CSScript.Evaluator.AutoResetEvaluatorOnError = true, this does not seem to be working.
This setting triggers so cold SoftReset, which is different to the Reset only by re-referencing the assemblies and re-creating the CompilerSettings object. It the time of the initial implementation SoftReset was sufficient to fully clear Mono.Evaluator. I will need to check may be it is not the case any more. I will let you know the outcome.
Where is the location of the cache folder, when CSScript.CacheEnabled is enabled?
The caching doesn't cover Mono Evaluator as the all assemblies are in the memory and cannot be cached.
Everything that is accessed through CSScript.Evaluator.* applicable to the Mono compiler and everything that it accessed through CSScript.* is a CodeDOM compiler, which indeed implements caching mechanism.
I am trying to use fixed point numbers in my VHDL project, but I keep having trouble implementing the library (found here http://www.eda-stds.org/fphdl/fixed_pkg_c.vhdl). The error I receive when trying to simulate is this
<ufixed> is not declared
My question is how exactly should a library be implemented so it can be used? As of now I have added it to the project in the IEEE_PROPOSED library, but it is not working. All source code can be found here https://github.com/srohrer32/beamformer/tree/fixed_num, under the hdl folder and libraries folder.
Are you using modelsim? Are you using a project? If not... I find the best way is to first compile the library on its own. Open your modelsim.ini file and make a path to the library. Like this:
lib_test = c:/test/source/lib_test
Finally, compile your own code and make sure you use the -modelsimini switch on vcom modelsim command.
If you are using a project (which I don't like, they are not as flexible) then you can point the project to the library.
More help about modelsim compiling with commands:
http://www.tkt.cs.tut.fi/tools/public/tutorials/mentor/modelsim/getting_started/gsms.html#compiling
Not being an isim user myself, a search through the ISim User Guide implies you need to create a separate project to compile into a library, contrasting with how easy it is to target a library from the command line.
Presumably you'd also need to add the library as a resource in your project. Funny there are no complaints about that yet you have:
library ieee_proposed;
in delay_calculation.vhd, noting that a library declaration simply makes the name available. Somewhere the implementation dependent mapping for the library name to library needs to be made. This by default is in xilinxisim.ini, but I imagine your project management interface allows you to map the library into your project, and isim should know where to look for the library.
Nosing around the user guide may be worthwhile.
In Simulation Steps Overview
User Libraries
Depending upon how you launch ISim, there are different methods
available to add user libraries:
When launching Project Navigator, define the user libraries in the ISE tool. See “Working with VHDL Libraries” in ISE Help for
details.
When using ISim standalone, interactive command mode, or non-interactive mode, set the library mapping file (see Appendix A,
Library Mapping File (xilinxisim.ini) to point to your logical or
physical libraries.
When launching ISim from the PlanAhead tool, define the user libraries in that tool. See the PlanAhead User Guide (UG632) for more
information. Appendix D, Additional Resources, contains a link to the
document.
See Working with VHDL Libraries, see To Create a VHDL Library and To Add Files to a VHDL Library.
(The top level link to ISE Help).
You'd think there'd be a FAQ for those of us apostate - speed reading 'religious' tomes sucks even using Google to find them. Notice the explanations are in terms of menu pull down actions, analogous to command line entry. We're being bitten by what's available on the top menu bar. And when you do manage to add and use a library successfully you'll remember how until someone changes the menus around, and you could of course wonder about documentation lagging.
Presumably what you've tried to do is set up the library mapping for synthesis mode in the ISE GUI, which is straightforward but completely ignored by iSim since it has its own system for managing library mappings. I'm not an iSim user but after looking at the documentation and a little testing it looks like the easiest way to set up a library is from the command line:
# This creates an ieee_proposed directory with a partially compiled object.
vhpcomp --work ieee_proposed=ieee_proposed fixed_pkg_c.vhdl
# Add a mapping from the logical library to the physical path.
# *nix shown. Windows would be similar or just use a text editor.
# <logical name>=<physical path>
echo ieee_proposed=`pwd`/ieee_proposed >> path/to/your/xilinxisim.ini
Make sure the xilinxisim.ini file is visible to iSim and it should pick up the mapping to your compiled library. You should be able to keep running vhpcomp from the parent of ieee_proposed to add more files to the library. You may have to manually copy the system default version to maintain the standard library mappings.
We have just started playing with Freemarker as UI template engine and the POC results are quite satisfactory and looks like we would start using it as our de facto UI processing framework instead of JSP.
The real benefit we see here we can store Freemarker templates anywhere (e.g. jar files, file system directories, even database for that matter !).
During POC, we did not see much performance issue but I was thinking if the templates can be pre-compiled and stored in file system / database for faster response on first hit or whenever there is a compiled cache-miss for a given template.
Is it possible to get hold of binary structure (most probably it would be multi-node tree) of a compiled Template and serialize in file and vice-versa.
Thanks in advance.
No, there's no such feature in FreeMarker (the AST had to be made serializable). OTOH first-time parsing shouldn't be slow unless you have a lot of templates and/or huge templates. By not slow I mean it's nothing like compiling JSP-s or Java which you can really "feel" when it first loads. Is there an actual performance problem, or it's just precaution?
I was wondering if there is a quick and effective way to remove all the unused variables (local, instance, even properties) in xcode... I am doing a code cleanup on my app and if I knew a quick way for code refactoring it would help me a lot...
Thanks...
It's being a long time since you made your question and maybe you found an answer already, but from an answer to a related question:
For static analysis, I strongly
recommend the Clang Static Analyzer
(which is happily built into Xcode 3.2
on Snow Leopard). Among all its other
virtues, this tool can trace code
paths an identify chunks of code that
cannot possibly be executed, and
should either be removed or the
surrounding code should be fixed so
that it can be called.
For dynamic analysis, I use gcov (with
unit testing) to identify which code
is actually executed. Coverage reports
(read with something like CoverStory)
reveal un-executed code, which —
coupled with manual examination and
testing — can help identify code that
may be dead. You do have to tweak some
setting and run gcov manually on your
binaries. I used this blog post to get
started.
Both methodologies are exactly for what you want, detecting unused code (both variables and methods) and removing them.
I am creating an [Iron]Ruby project that needs to support several environments (more specifically, WPF, Silverlight, and WinForms - but that's not as important). The following is an actual, concrete example of where I'm stuck:
I have to implement a Bitmap class as part of a library, and this class will need to be implemented differently depending on what environment it's running in (e.g. if I'm running this in the browser as a silverlight app, I won't have access to methods that would be available on the desktop). And here's the catch - I don't control the instantiation of Bitmap, nor any of the other classes within the library. Why? Because it's a port of another application; and, while I do have the code for the application, I don't want to break compatibility by changing that code. I do, however, control the entry point to the application, so I can require whatever I need, perform setup, configure global variables, etc.
Edit: If you're curious, this is the project I'm working on:
http://github.com/cstrahan/open-rpg-maker
Here's what I want to know:
How should I set the configuration at startup, such that Bitmap will behave appropriately?
How should I structure this in my git repo / source tree?
Here are some of my thoughts, but I'm sure you'll have better ideas:
How should I set the configuration at startup?
When distributing the app, place a require at the top depending on the targeted environment, like so: require silverlight/bitmap. In this case, lib/bitmap.rb would be empty, while lib/silverlight/bitmap.rb would contain the implementation. Or...
Stuff all implementations in lib/bitmap.rb, and conditionally execute based on a class instance variable or constant: Bitmap.impl = "silverlight". Or...
Maintain a separate branch for each distro - despite the library being almost exactly the same.
How should I structure this in my git repo / source tree?
Separate branches per distribution. Or...
Separate implementation-specific subfolders (e.g. lib/silverlight/bitmap.rb).
Being very new to Ruby, I'm not very familiar with such best practices (I'm coming from C#). Any advice would be greatly appreciated!
-Charles