Bartosz Taudul
0da1e8551f
Track lock contention status.
2019-05-12 16:17:17 +02:00
Bartosz Taudul
4850e19ebd
Store color in message data.
2019-05-10 20:26:27 +02:00
Bartosz Taudul
a7886cf82c
Replace linear search with hash lookup.
2019-04-03 16:24:16 +02:00
Bartosz Taudul
7e6a8135df
Remove double indirection in GetNextLockEvent().
2019-03-16 14:18:43 +01:00
Bartosz Taudul
4d66317bc3
Add per-thread time ranges to lock maps.
2019-03-16 02:50:51 +01:00
Bartosz Taudul
9fc022346b
Replace frame pointers with callstack frame ids.
2019-03-03 18:05:03 +01:00
Bartosz Taudul
1feedb17ac
Add callstack frame identifier and the required plumbing.
2019-03-03 18:05:03 +01:00
Bartosz Taudul
e190faa7e1
Save/load CPU usage plot.
2019-02-21 22:56:59 +01:00
Bartosz Taudul
e9baa80bf3
Process CPU usage reports.
2019-02-21 22:56:59 +01:00
Bartosz Taudul
b945f83169
Don't separate inclusive/exclusive counts.
...
There is no way for one frame to have both. Coloring is preserved and is
now determined by presence of children.
2019-02-06 22:36:21 +01:00
Bartosz Taudul
ddad475c19
Make it possible to store multiple frames at single frame address.
2019-01-20 19:11:48 +01:00
Bartosz Taudul
dcc6bee607
Process zone validation messages.
2019-01-14 22:56:10 +01:00
Bartosz Taudul
9360df89b1
Store announce and terminate time of locks.
2018-12-16 21:07:26 +01:00
Bartosz Taudul
9301986bae
Collect callstacks for each entry in call stack tree.
2018-09-27 22:56:44 +02:00
Bartosz Taudul
3b526b074e
Send crash report.
2018-08-20 02:23:55 +02:00
Bartosz Taudul
df14cf5330
Implement callstack tree of memory allocations.
2018-08-14 18:37:06 +02:00
Bartosz Taudul
9d051cf5ee
Add support for discontinuous frames.
2018-08-05 02:15:54 +02:00
Bartosz Taudul
23dfc2e3fc
Multiple frame sets support.
2018-08-04 21:10:45 +02:00
Bartosz Taudul
7d7877517e
Also remove child vectors from GPU events.
2018-07-22 19:47:01 +02:00
Bartosz Taudul
3a934b2ba3
Store children vectors in a separate data collection.
...
This reduces per-zone memory cost by 9 bytes if there are no children
and increases it by 4 bytes, if there are children. This is universally
a better solution, as the following data shows:
+++ /home/wolf/desktop/tracy-old/android.tracy +++
Vectors: 2794480
Size 0: 2373070 (84.92%)
Size 1: 70237 (2.51%)
Size 2+: 351173 (12.57%)
+++ /home/wolf/desktop/tracy-old/asset-new.tracy +++
Vectors: 1799227
Size 0: 1482691 (82.41%)
Size 1: 93272 (5.18%)
Size 2+: 223264 (12.41%)
+++ /home/wolf/desktop/tracy-old/asset-new-id.tracy +++
Vectors: 1977996
Size 0: 1640817 (82.95%)
Size 1: 97198 (4.91%)
Size 2+: 239981 (12.13%)
+++ /home/wolf/desktop/tracy-old/asset-old.tracy +++
Vectors: 1782395
Size 0: 1471437 (82.55%)
Size 1: 88813 (4.98%)
Size 2+: 222145 (12.46%)
+++ /home/wolf/desktop/tracy-old/big.tracy +++
Vectors: 180794047
Size 0: 172696094 (95.52%)
Size 1: 2799772 (1.55%)
Size 2+: 5298181 (2.93%)
+++ /home/wolf/desktop/tracy-old/darkrl.tracy +++
Vectors: 12014129
Size 0: 11611324 (96.65%)
Size 1: 134980 (1.12%)
Size 2+: 267825 (2.23%)
+++ /home/wolf/desktop/tracy-old/mem.tracy +++
Vectors: 383097
Size 0: 321932 (84.03%)
Size 1: 854 (0.22%)
Size 2+: 60311 (15.74%)
+++ /home/wolf/desktop/tracy-old/new.tracy +++
Vectors: 77536
Size 0: 63035 (81.30%)
Size 1: 8886 (11.46%)
Size 2+: 5615 (7.24%)
+++ /home/wolf/desktop/tracy-old/selfprofile.tracy +++
Vectors: 22940871
Size 0: 22704868 (98.97%)
Size 1: 73000 (0.32%)
Size 2+: 163003 (0.71%)
+++ /home/wolf/desktop/tracy-old/tbrowser.tracy +++
Vectors: 962682
Size 0: 695380 (72.23%)
Size 1: 43007 (4.47%)
Size 2+: 224295 (23.30%)
+++ /home/wolf/desktop/tracy-old/virtualfile_hc.tracy +++
Vectors: 529170
Size 0: 449386 (84.92%)
Size 1: 15694 (2.97%)
Size 2+: 64090 (12.11%)
+++ /home/wolf/desktop/tracy-old/zfile_hc.tracy +++
Vectors: 264849
Size 0: 220589 (83.29%)
Size 1: 9386 (3.54%)
Size 2+: 34874 (13.17%)
2018-07-22 16:05:50 +02:00
Bartosz Taudul
053284b1c7
Process custom free-form zone names.
2018-06-29 16:12:17 +02:00
Bartosz Taudul
4a467b6d03
Remove GPU resync leftovers.
2018-06-28 00:48:23 +02:00
Bartosz Taudul
11cf650be6
Fix GPU queries ordering.
...
With multithreaded Vulkan rendering it is possible that GPU time queries
will be sent in a different order than the originating CPU queries were
made. This commit changes the in-order queue to a map of queries,
waiting to be resolved.
2018-06-22 16:37:54 +02:00
Bartosz Taudul
35dc2f796e
Process GpuZoneBeginCallstack queue event.
2018-06-22 01:56:32 +02:00
Bartosz Taudul
205a4e4ca2
Add callstack index to ZoneEvent.
2018-06-22 01:11:03 +02:00
Bartosz Taudul
2a618c90d5
Properly save compressed thread in GPU events.
2018-06-20 23:12:49 +02:00
Bartosz Taudul
88b1955a5a
Filename in callstack frame is not a persistent pointer.
2018-06-20 01:26:05 +02:00
Bartosz Taudul
203744cdd9
Callstack frame queries.
2018-06-20 00:25:26 +02:00
Bartosz Taudul
06f34052a5
Have to track callstacks of both alloc and free.
2018-06-19 22:08:47 +02:00
Bartosz Taudul
e03493f082
Store callstack index as uint32_t.
2018-06-19 21:39:22 +02:00
Bartosz Taudul
59dc55002b
Callstack ptr in server data structures.
...
Will be probably reduced to 32-bit index later on.
2018-06-19 18:52:10 +02:00
Bartosz Taudul
bb0631585c
Store thread id of GPU events.
2018-06-17 19:07:07 +02:00
Bartosz Taudul
dcd6cac078
Save GPU timestamp period.
...
Bump file version to 0.3.2.
2018-06-17 18:27:42 +02:00
Bartosz Taudul
53aea660c8
Store thread id in MessageData.
2018-05-25 21:10:38 +02:00
Bartosz Taudul
b18841aa75
Store ordered list of memory frees.
2018-05-02 17:59:50 +02:00
Bartosz Taudul
bc84ebc338
Read/write LockEvent data in one go.
2018-04-29 03:41:58 +02:00
Bartosz Taudul
9769cc4d7d
Read/write most of MemEvent in one go.
2018-04-29 03:41:58 +02:00
Bartosz Taudul
d8bfe7de2e
Create memory plot based on memory alloc/free events.
2018-04-28 15:49:12 +02:00
Bartosz Taudul
cd34ed6968
Two plot types: user and memory.
...
Only user plots are saved in a dump file.
2018-04-28 15:48:05 +02:00
Bartosz Taudul
bf99bff87d
Store MemEvents directly in the vector.
2018-04-03 14:17:51 +02:00
Bartosz Taudul
52f59c90bf
Track memory usage.
2018-04-02 00:00:49 +02:00
Bartosz Taudul
a574f98f0c
Memory events are now serialized.
2018-04-01 20:13:01 +02:00
Bartosz Taudul
b12375815c
Broken memory events processing.
2018-04-01 02:03:34 +02:00
Bartosz Taudul
fe6c753f12
Store lock thread map in flat hash map.
2018-03-20 15:40:25 +01:00
Bartosz Taudul
9dfa9c95cb
Read and write whole ZoneEvent/GpuEvent data at once.
2018-03-15 21:59:16 +01:00
Bartosz Taudul
5cb917e868
No nonsense union.
2018-03-04 17:52:51 +01:00
Bartosz Taudul
5afdccfc46
Properly initialize data.
...
Unused bitbield bits and inactive string index/reference had thrash
values in release builds, which prevented de-duplication of source
location payloads.
2018-03-04 17:47:26 +01:00
Bartosz Szreder
9e3f18a62a
Split data handling code from the view.
2018-02-21 16:41:37 +01:00
Bartosz Szreder
d5fe006e2d
Add missing include charutil::hash()
2018-02-12 19:07:55 +01:00
Bartosz Taudul
d0d3bb1141
Store shared lock bits only for shared locks.
2017-12-17 18:44:31 +01:00
Bartosz Taudul
340506406e
Shared lock state machine.
2017-12-10 23:30:13 +01:00
Bartosz Taudul
b07718ab9e
Track list of shared locks.
2017-12-10 22:42:39 +01:00
Bartosz Taudul
398eecbb94
Store LockEvent type as an enum class.
2017-12-10 22:37:56 +01:00
Bartosz Taudul
3567d7edd8
Reintroduce lock announce events.
2017-12-10 21:40:48 +01:00
Bartosz Taudul
981bbfe42d
Reorder LockEvent fields.
2017-12-09 19:13:59 +01:00
Bartosz Taudul
48678b3bd7
Drop bitfield usage.
2017-12-05 22:34:48 +01:00
Bartosz Taudul
081087b9ce
Drop an indirection level in plots.
2017-12-05 21:24:09 +01:00
Bartosz Taudul
5246098c79
GPU context hiding plumbing.
2017-11-30 15:31:31 +01:00
Bartosz Taudul
a515bf8878
Perform GPU to CPU resynchronization on each collect event.
2017-11-25 13:33:57 +01:00
Bartosz Taudul
f2a19c4593
Use flat_hash_map for source location payload map.
2017-11-24 01:10:12 +01:00
Bartosz Taudul
0fe4232015
Optimize string wrappers initialization.
2017-11-19 13:36:03 +01:00
Bartosz Taudul
b3e0d2660d
Properly use zone stack when inserting new zones.
...
CPU zones stack was also moved to thread data struct.
2017-11-19 01:16:21 +01:00
Bartosz Taudul
123de74fca
Thread thread zone count.
2017-11-18 01:14:16 +01:00
Bartosz Taudul
cc3d18d51d
Count GPU zones.
2017-11-18 01:07:28 +01:00
Bartosz Taudul
1a4889116e
Store GPU accuracy bits info.
2017-11-18 00:34:16 +01:00
Bartosz Taudul
515443605d
memset on non-POD data is UB.
2017-11-16 02:19:52 +01:00
Bartosz Taudul
a3ba188af5
Manual initialization of GpuEvents.
2017-11-15 22:21:02 +01:00
Bartosz Taudul
afa9eec5dd
Manual ZoneEvent vector initialization.
2017-11-15 22:13:23 +01:00
Bartosz Taudul
3a60a28a71
Don't allocate unneeded memory.
2017-11-15 21:26:58 +01:00
Bartosz Taudul
7187e1e5f5
Some strings are always indices.
...
This saves 4 bytes.
2017-11-15 10:56:27 +01:00
Bartosz Taudul
8b2af369ad
Store zone text in StringRef. Use name from source location.
2017-11-15 10:56:27 +01:00
Bartosz Taudul
c43eb29ce0
Don't send source location pointer in query reply.
...
Since reply order is the same as the query order, the server already
knows what source location it receives. This observation allows placing
zone name into the source location struct.
2017-11-14 23:06:45 +01:00
Bartosz Taudul
5c872b2137
Simplify GPU context handling.
2017-11-14 00:48:26 +01:00
Bartosz Taudul
af81d999e9
Draw GPU zones.
2017-11-11 22:56:05 +01:00
Bartosz Taudul
6fcdb924e8
CPU-side GPU event transfer.
2017-11-11 21:09:48 +01:00
Bartosz Taudul
3c00ce0958
GPU context registration.
2017-11-11 19:44:09 +01:00
Bartosz Taudul
ce35009c63
Move structs out of TracyView.
2017-11-11 19:21:25 +01:00
Bartosz Taudul
81735aea2f
Support for setting zone names in lua.
2017-11-11 17:56:41 +01:00
Bartosz Taudul
ca4483ecf5
Prevent source location payload duplication.
2017-11-11 02:31:51 +01:00
Bartosz Taudul
24084cbcd2
Standard way of string reference storage in SourceLocation.
...
StringRef::isptr was changed to isidx, as initialization of empty
SourceLocation zeroes the struct.
2017-11-11 02:02:47 +01:00
Bartosz Taudul
947cd04e5e
General solution for string pointer/index problem.
2017-11-11 01:39:34 +01:00
Bartosz Taudul
c6a7bcb086
Store source location payloads.
...
No saving yet. No detection of duplicate entries.
2017-11-05 21:24:50 +01:00
Bartosz Taudul
3d2450fc10
Store source locations in a proper data structure.
2017-11-05 20:54:49 +01:00
Bartosz Taudul
9524b6447e
Pack LockEvent data, saving one byte.
2017-10-29 16:49:22 +01:00
Bartosz Taudul
39ac41bc19
Store source location on 32 bits, not 64.
2017-10-28 22:14:01 +02:00
Bartosz Taudul
cc8683a399
Store TextData pointer as an index in array.
...
This further reduces ZoneEvent size by 4 bytes.
2017-10-22 16:40:15 +02:00
Bartosz Taudul
2e6350877d
Remove parent pointer from ZoneEvent.
2017-10-22 16:15:27 +02:00
Bartosz Taudul
7345ca5aa6
Rename Event to ZoneEvent.
2017-10-22 15:37:24 +02:00
Bartosz Taudul
75457c1465
Remove +x flag from files.
2017-10-10 21:56:15 +02:00
Bartosz Taudul
a6c5993401
Fix locks with more than two threads.
2017-10-08 23:03:38 +02:00
Bartosz Taudul
71e16d069e
Store lock mark locations.
2017-10-06 18:04:17 +02:00
Bartosz Taudul
78067eb35e
Calculate lock wait counts.
2017-10-05 14:02:08 +02:00
Bartosz Taudul
4c8e9f7d5d
Track number of held locks.
2017-10-04 19:42:44 +02:00
Bartosz Taudul
54735bacea
Lock events dispatch.
2017-10-04 18:51:50 +02:00
Bartosz Taudul
c2bccf7126
Move towards proper data structures.
2017-10-04 18:32:53 +02:00
Bartosz Taudul
4fad4be816
Pack event structures.
2017-10-04 18:12:11 +02:00
Bartosz Taudul
78f8425dc7
Announce lock creation.
2017-10-04 16:16:40 +02:00
Bartosz Taudul
91fd171b60
Store CPU id in zones.
2017-10-01 19:17:29 +02:00
Bartosz Taudul
8c1c395cec
Allow sending custom zone names.
2017-09-28 19:28:24 +02:00
Bartosz Taudul
91e6210e34
Store zone text data in a separate structure.
2017-09-28 19:06:39 +02:00
Bartosz Taudul
d1bbb731fc
Zone text (custom string) transfer.
2017-09-27 02:18:17 +02:00
Bartosz Taudul
e90a86e06e
Store zone color in source location struct.
2017-09-26 18:54:48 +02:00
Bartosz Taudul
7424077d70
Store source location in a single object.
...
Source file, function name and line number are now stored in a const
static container object. This has the following benefits:
- Slightly lighter profiling workload (3 instructions less).
- Profiling queue event size is significantly reduced, by 12 bytes. This
has an effect on all queue event types.
- Source location grouping has now no cost, as it's performed at the
compilation stage. This allows simplification of server code.
The downside is that the full source location resolution is now
performed in two steps, as the server has to query both source location
container and strings contained within. This has almost no real impact
on profiler operation.
2017-09-26 02:39:08 +02:00
Bartosz Taudul
519cb8dff3
Allow adding custom colors to zones.
2017-09-25 22:46:14 +02:00
Bartosz Taudul
a68b0921b1
Add parent pointer to Event.
2017-09-23 15:41:26 +02:00
Bartosz Taudul
5065743bf0
Store zone source location.
2017-09-21 21:57:40 +02:00
Bartosz Taudul
b4faa0a9b9
Use small vector.
2017-09-15 20:17:39 +02:00
Bartosz Taudul
1c56347f1d
Use slab allocator to store event data.
2017-09-15 19:56:55 +02:00
Bartosz Taudul
a159b70d40
Reconstruct event data.
2017-09-14 02:00:13 +02:00