When debugging gem5, I want to know what microops are in the reorder buffer, instruction queue or decoding queue? Is there any way to do it?
When debugging with gdb, when I use p queuename, the display is similar to
{std::queue wrapping: std::deque with 8 elements = {{data = 0x23e5d700}, {data = 0x23ead180}, {
data = 0x23e722c0}, {data = 0x2460a580}, {data = 0x37ecc00}, {data = 0x23ef2580}, {data = 0x23eb8ec0}, {
data = 0x23ec2b00}}, std::queue wrapping: std::deque with 0 elements,
std::queue wrapping: std::deque with 0 elements, std::queue wrapping: std::deque with 0 elements}
But I want it to show specific micro-operations such as mov add.
Related
I am rendering a single mesh test object that contains 8000 cubes, 1 monkey, and 1 plane. I checked normals and winding orders. all seems correct for mesh.
if I render a single instance, the depth buffer works correctly. if I render using instanced when instant number increase depth buffer doesn't work correctly.
//Depth Attachment
//Create Depth Image
depthImage = new Image(
physicalDevice_,
logicalDevice_,
VK_IMAGE_TYPE_2D,
depthFormat,
VK_IMAGE_TILING_OPTIMAL,
swapChainExtent.width,
swapChainExtent.height,
1,
sampleCount_, //do we really need to sample depth????
VK_IMAGE_LAYOUT_UNDEFINED,
VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT,
VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT);
depthImageView = new ImageView(
logicalDevice_,
VK_IMAGE_VIEW_TYPE_2D,
depthImage->handle,
depthFormat,
1,
VK_IMAGE_ASPECT_DEPTH_BIT);
//Render pass Depth Attachment
VkAttachmentDescription depthAttachment{};
depthAttachment.format = depthFormat;
depthAttachment.samples = VK_SAMPLE_COUNT_1_BIT;
depthAttachment.loadOp = VK_ATTACHMENT_LOAD_OP_CLEAR;
depthAttachment.storeOp = VK_ATTACHMENT_STORE_OP_DONT_CARE;
depthAttachment.stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE;
depthAttachment.stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE;
depthAttachment.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED;
depthAttachment.finalLayout = VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL;
//Graphics Pipeline
VkPipelineDepthStencilStateCreateInfo depthStencil{};
depthStencil.sType = VK_STRUCTURE_TYPE_PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO;
depthStencil.depthTestEnable = VK_TRUE;
depthStencil.depthWriteEnable = VK_TRUE;
depthStencil.depthCompareOp = VK_COMPARE_OP_LESS;
depthStencil.depthBoundsTestEnable = VK_FALSE;
depthStencil.minDepthBounds = 0.0f; // Optional
depthStencil.maxDepthBounds = 1.0f; // Optional
depthStencil.stencilTestEnable = VK_FALSE;
depthStencil.front = {}; // Optional
depthStencil.back = {}; // Optional
Below the screen is the view from the 100th instance of that mesh.
The higher the instance number, the more the depth buffer deteriorates.
The result is still the same if I use a different model.
What am I missing?
Result:
Depth Buffer:
For those who may struggle with similar issue. I found a solution.
Actually, this issue is not related to instancing directly but it's a bit confusing issue becomes visible as the instance index increases.
the cause is projection matrix z-near value was so small.
...
fieldOfView = 65.0f;
znear = 0.001f; // ! this much small value cause this problem
zfar = 5000.0f;
...
changed z-near to an acceptable value. Problem solved.
...
fieldOfView = 65.0f;
znear = 0.1f;
zfar = 5000.0f;
...
I am trying to mimic the autosave function in GMS v3 so that I can use in version 1 and 2. I would like to first acknowledge that the main bulk of the script originates from Dr Bernhard Schaffer's "How to script... Digital Micrograph Scripting Handbook". I have modified it a bit, so that any new image recorded by the camera can be autosave into the file. However, I met some problems because if I decide to click on live-view image and move the image around, or using live-fft, the live view image or the FFT image will be saved as well. One of the ideas I have is to use the taggroup information such as the "Acquisition:Parameters:Parameter Set Name" because for live view or live-FFT, this would be either in search or focus mode. Another idea is to use the document ID e.g iDocID = idoc.ImageDocumentGETID() to locate the ID of the live image. However, i am clueless then how to use this information to exclude them from autosaving. Can anyone point to me how i can proceed with this script?
Below is the script
Class PeriodicAutoSave : Object
{
Number output
PeriodicAutoSave(Object self) Result("\n Object ID"+self.ScriptObjectGetID()+" created.")
~PeriodicAutoSave(Object self) Result("\n Object ID"+self.ScriptObjectGetID()+" destroyed")
Void Init2(Object self, Number op)
output=op
Void AutoSave_SaveAll(Object self)
{
String path, name, targettype, targettype1, area, mag, mode, search, result1
ImageDocument idoc
Number nr_idoc, count, index_i, index, iDocID, iDocID_search
path = "c:\\path\\"
name = "test"
targettype=="Gatan Format (*.dm4)"
targettype1 = "dm4"
If (output) Result("\n AutoSave...")
nr_idoc = CountImageDocuments()
For (count = 1; count<nr_idoc; count++)
{
idoc = GetImageDocument(count) //imagedocument
index = 1 // user decide the index to start with
index_i= nr_idoc - index
If (idoc.ImageDocumentIsDirty())
{
idoc = getfrontimagedocument()
iDocID = idoc.ImageDocumentGetID()
TagGroup tg = ImageGetTagGroup(idoc.ImageDocumentGetImage(0)) // cannot declare an 'img' for this line as it will prompt an error?
tg.TagGroupGetTagAsString("Microscope Info:Formatted Indicated Mag", mag)
Try{
{
idoc.ImageDocumentSavetoFile( "Gatan Format", path+index_i+"-"+name+"-"+mag+".dm4")
idoc.ImageDocumentSetName(index_i + "-"+name+"-"+mag+".dm4")
idoc.ImageDocumentClean()
}
If (Output) Result("\n\t saving: "+idoc.ImageDocumentGetCurrentFile())
}
Catch{
Result("\n image cannot be saved at the moment:" + GetExceptionString())
Break
}
Result("\ Continue autosave...")
}
}
}
}
Void LaunchAutoSave()
{
Object obj = Alloc(PeriodicAutoSave)
obj.Init2(2)
Number task_id = obj.AddMainThreadPeriodicTask("AutoSave_SaveALL",6)
//Sleep(10)
while(!shiftdown()) 1==2
RemoveMainThreadTask(task_id)
}
LaunchAutoSave()
thank you very much for your pointers! I have tried and it works very well with my script. as the 'TagGroupDoesTagExist' only refers to the taggroup, I modified further to include the tags I want to filter e.g "Search" or "Focus" and it seems to work well. The script that I modified to your existing ones is as below :
If (idoc.ImageDocumentIsDirty())
{
//now find out what is a filter condition and skip if it is true
skip = 0
TagGroup tg = idoc.ImageDocumentGetImage(0).ImageGetTagGroup()
tg.TagGroupGetTagAsString("Microscope Info:Formatted Indicated Mag", mag)
tg.TagGroupGetTagAsString("Acquisition:Parameters:Parameter Set Name", mode)
skip = tg.TagGroupDoesTagExist("Acquisition:Parameters:Parameter Set Name")
if(skip && (mode == "Search" || mode== "Focus")) continue
Your idea of filtering is a good one, but there is something strange with your for loop.
in
nr_idoc = CountImageDocuments()
For (count = 1; count<nr_idoc; count++)
{
idoc = GetImageDocument(count) //imagedocument
you iterate over all currently open imageDocuments (except the first one!?) and get them one by one, but then in
If (idoc.ImageDocumentIsDirty())
{
idoc = getfrontimagedocument()
you actually get the front-most (selected) document instead each time. Why are you doing this?
Why not go with:
number nr_idoc = CountImageDocuments()
for (number count = 0; count<nr_idoc; count++)
{
imagedocument idoc = GetImageDocument(count)
If (idoc.ImageDocumentIsDirty())
{
// now find out what is a filter condition and skip if it is true
number skip = 0
TagGroup tg = idoc.ImageDocumentGetImage(0).ImageGetTagGroup()
skip = tg.TagGroupDoesTagExist("Microscope Info:Formatted Indicated Mag")
if (skip) continue
// do saving
}
}
I want to record audio with AudioKit. What I want to achieve is a buffer with 512 bytes if the buffer is full. It will translate the buffer info through TCP. And the current situation is that I only know how to restore the Audio data into a file like in the code below.
Can anyone provide some tips to achieve my goal?
Here is my code:
AKAudioFile.cleanTempDirectory()
AKSettings.bufferLength = .medium
do {
try AKSettings.setSession(category: .playAndRecord, with: .allowBluetoothA2DP)
} catch {
AKLog("Could not set session category.")
}
AKSettings.defaultToSpeaker = true
micMixer = AKMixer(mic)
micBooster = AKBooster(micMixer)
micBooster.gain = 0
recorder = try?AKNodeRecorder.init(node: micMixer)
if let file = recorder.audioFile {
print(file.directoryPath)
player = AKPlayer(audioFile: file)
}
player.isLooping = true
moogLadder = AKMoogLadder(player)
mainMixer = AKMixer(moogLadder, micBooster)
AudioKit.output = mainMixer
let clipRecoder = AKClipRecorder(node: mainMixer)
do {
try AudioKit.start()
} catch {
AKLog("AudioKit did not start!")
}
try! recorder.record()
When listening to a long string output from a shell process, I receive the data in two chunks. How can I get the entire text?
Here is the code in question:
int i = 0;
Process.start('perl', ['print_text.pl']).then((Process p) {
p.stdout.transform(UTF8.decoder).listen((data) => print("${i++} ${data}"));
p.stdin.writeln('print');
});
The result from running this code is:
0 text.....
1 text.....
I've reported this issue as a bug here. You can run the sample app attached to the report to see the issue.
Try using UTF8.decodeStream().
import 'dart:io';
import 'dart:convert' show UTF8;
main() {
Process.start('ls', ['-la'])
.then((p) => UTF8.decodeStream(p.stdout))
.then((s) => print('Output:\n$s'));
}
I solved this problem by doing the following:
In the shell script, tag the start and the end of the shell's output with a random number.
In Dart, concatenate all the chunks together and use the tags to check if you got the complete result.
This solution does not require exiting the process to obtain the process result.
This is a sample code:
String text = '';
process.stdout.transform(UTF8.decoder).listen((String chunk) {
text = text + chunk;
if (text.substring(errors.length - 1) == text.substring(0, 1)) {
text = text.replaceFirst(new RegExp(r'^(\d+)'), '').replaceFirst(new RegExp(r'(\d+)$'), '');
// use process result then clear text value for subsequent process results
text = '';
}
});
It's not clear what the actual question is but I assume this is what you are looking for:
import 'dart:io';
import 'dart:convert' show UTF8;
void main() {
int i = 0;
String text = '';
Process.start('perl', ['analyze.pl']).then((p) {
p.stdout.transform(UTF8.decoder).listen((data) =>
text += data);
p.exitCode.then((_) => print(text));
});
}
I am testing flume to load data into hHase and thinking about parallel data loading with using flume's selector and inteceptor, because of speed gap between source and sink.
So, what I want to do with flume are
creating Event's header with interceptors's regex_extractor type
multiplexing Event with header to more than two channels with selector's multiplexing type
in one source-channel-sink.
and tried configuration as below.
agent.sources = tailsrc
agent.channels = mem1 mem2
agent.sinks = std1 std2
agent.sources.tailsrc.type = exec
agent.sources.tailsrc.command = tail -F /home/flumeuser/test/in.txt
agent.sources.tailsrc.batchSize = 1
agent.sources.tailsrc.interceptors = i1
agent.sources.tailsrc.interceptors.i1.type = regex_extractor
agent.sources.tailsrc.interceptors.i1.regex = ^(\\d)
agent.sources.tailsrc.interceptors.i1.serializers = t1
agent.sources.tailsrc.interceptors.i1.serializers.t1.name = type
agent.sources.tailsrc.selector.type = multiplexing
agent.sources.tailsrc.selector.header = type
agent.sources.tailsrc.selector.mapping.1 = mem1
agent.sources.tailsrc.selector.mapping.2 = mem2
agent.sinks.std1.type = file_roll
agent.sinks.std1.channel = mem1
agent.sinks.std1.batchSize = 1
agent.sinks.std1.sink.directory = /var/log/flumeout/1
agent.sinks.std1.rollInterval = 0
agent.sinks.std2.type = file_roll
agent.sinks.std2.channel = mem2
agent.sinks.std2.batchSize = 1
agent.sinks.std2.sink.directory = /var/log/flumeout/2
agent.sinks.std2.rollInterval = 0
agent.channels.mem1.type = memory
agent.channels.mem1.capacity = 100
agent.channels.mem2.type = memory
agent.channels.mem2.capacity = 100
But, it doesn't work!
when selector part is removed, there are some interceptor debugging message in flume's log.
but when selector and interceptor are together, there are nothing.
Is there any wrong expression or something I missed?
Thanks for reading. :)
I found it.
In the flume log, there are warning message as below.
2013-10-10 16:34:20,514 (conf-file-poller-0) [WARN - org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.validateSources(FlumeConfiguration.java:571)] Removed tailsrc due to Failed to configure component!
so I had attached below line
agent.sources.tailsrc.channels = mem1 mem2
and then It works!!!!