Merging the Raycaster with the Sampler — Handmade Hero — Episode Guide

0:04Recap and set the stage for the day

🗩

0:04Recap and set the stage for the day

🗩

0:04Recap and set the stage for the day

🗩

0:25Our world remains in the dark

🏃

0:25Our world remains in the dark

🏃

0:25Our world remains in the dark

🏃

1:08Let GridRayCast() set a non-zero CostMetric

1:25Demo the current lighting

🏃

1:25Demo the current lighting

🏃

1:25Demo the current lighting

🏃

1:36Describe and consider the performance of our lighting

🗩

1:36Describe and consider the performance of our lighting

🗩

1:36Describe and consider the performance of our lighting

🗩

4:39Rare 4coder crash

🗹

4:39Rare 4coder crash

🗹

4:39Rare 4coder crash

🗹

5:26Consider gauging the grid ray casting performance if ComputeVoxelIrradianceAt() was optimal

🗩

5:26Consider gauging the grid ray casting performance if ComputeVoxelIrradianceAt() was optimal

🗩

5:26Consider gauging the grid ray casting performance if ComputeVoxelIrradianceAt() was optimal

🗩

8:25Break into ComputeVoxelIrradianceAt()

🏃

8:25Break into ComputeVoxelIrradianceAt()

🏃

8:25Break into ComputeVoxelIrradianceAt()

🏃

9:29RemedyBG feature request: Tabulated / colourised disassembly

🗹

9:29RemedyBG feature request: Tabulated / colourised disassembly

🗹

9:29RemedyBG feature request: Tabulated / colourised disassembly

🗹

12:17Inspect the assembly of ComputeVoxelIrradianceAt()

🏃

12:17Inspect the assembly of ComputeVoxelIrradianceAt()

🏃

12:17Inspect the assembly of ComputeVoxelIrradianceAt()

🏃

12:26Research the comiss instruction¹^,2

📖

12:26Research the comiss instruction¹^,2

📖

12:26Research the comiss instruction¹^,2

📖

15:40Try to interpret the origin of the comiss instructions

📖

15:40Try to interpret the origin of the comiss instructions

📖

15:40Try to interpret the origin of the comiss instructions

📖

21:06Compare our f32_4x and f32 versions of AbsoluteValue() in the Compiler Explorer³

📖

21:06Compare our f32_4x and f32 versions of AbsoluteValue() in the Compiler Explorer³

📖

21:06Compare our f32_4x and f32 versions of AbsoluteValue() in the Compiler Explorer³

📖

29:58Point out our comiss instructions

🏃

29:58Point out our comiss instructions

🏃

29:58Point out our comiss instructions

🏃

30:16Redo our f32 version of AbsoluteValue() based on the f32_4x version⁴^,5

33:09Our comiss instructions are replaced with a call, which in turn retains the comiss ones

🏃

33:09Our comiss instructions are replaced with a call, which in turn retains the comiss ones

🏃

33:09Our comiss instructions are replaced with a call, which in turn retains the comiss ones

🏃

34:15Revert the f32 version of AbsoluteValue() to use fabs()

34:27Our comiss instructions are back, welded in

🏃

34:27Our comiss instructions are back, welded in

🏃

34:27Our comiss instructions are back, welded in

🏃

34:47Weld GetOctahedralOffset() in to ComputeVoxelIrradianceAt()

36:37Our comiss instructions remain

🏃

36:37Our comiss instructions remain

🏃

36:37Our comiss instructions remain

🏃

38:01Weld OctahedralFromUnitVector() in to ComputeVoxelIrradianceAt()

39:10SignOf() is the source of two comiss instructions

📖

39:10SignOf() is the source of two comiss instructions

📖

39:10SignOf() is the source of two comiss instructions

📖

40:04Make SignOf() branchless⁶

44:34Two of our comiss instructions are gone

🏃

44:34Two of our comiss instructions are gone

🏃

44:34Two of our comiss instructions are gone

🏃

45:26The lighting still looks the same

🏃

45:26The lighting still looks the same

🏃

45:26The lighting still looks the same

🏃

46:12Make the UV computation in ComputeVoxelIrradianceAt() fully branchless

50:06Note why _mm_extract_ps() is often not a good idea, as mmozeiko pointed out⁷

🗩

50:06Note why _mm_extract_ps() is often not a good idea, as mmozeiko pointed out⁷

🗩

50:06Note why _mm_extract_ps() is often not a good idea, as mmozeiko pointed out⁷

🗩

53:18Finish making the UV computation in ComputeVoxelIrradianceAt() fully branchless, also noting to change Extract1() and Extract2() to use SHUF+CVTSS

54:16All of our comiss instructions are gone

🏃

54:16All of our comiss instructions are gone

🏃

54:16All of our comiss instructions are gone

🏃

54:46Seek improvements to our ComputeVoxelIrradianceAt() vectorisation

📖

54:46Seek improvements to our ComputeVoxelIrradianceAt() vectorisation

📖

54:46Seek improvements to our ComputeVoxelIrradianceAt() vectorisation

📖

59:54Remove the stupidity from our ComputeVoxelIrradianceAt() vectorisation⁸

1:05:02A few words on being aware of gotchas in poorly designed instruction sets

🗩

1:05:02A few words on being aware of gotchas in poorly designed instruction sets

🗩

1:05:02A few words on being aware of gotchas in poorly designed instruction sets

🗩

1:06:45Continue to remove the stupidity from our ComputeVoxelIrradianceAt() vectorisation, introducing an f32_4x version of SignOf()

1:12:38Hit a read access violation in ComputeVoxelIrradianceAt()

🏃

1:12:38Hit a read access violation in ComputeVoxelIrradianceAt()

🏃

1:12:38Hit a read access violation in ComputeVoxelIrradianceAt()

🏃

1:13:38Remove the problematic part of the BaseXYZ computation in ComputeVoxelIrradianceAt()

1:13:50We no longer hit that read access violation

🏃

1:13:50We no longer hit that read access violation

🏃

1:13:50We no longer hit that read access violation

🏃

1:14:08Scrutinise ComputeVoxelIrradianceAt() for bugs

📖

1:14:08Scrutinise ComputeVoxelIrradianceAt() for bugs

📖

1:14:08Scrutinise ComputeVoxelIrradianceAt() for bugs

📖

1:21:25Make ComputeVoxelIrradianceAt() call GetOctahedralOffset() as originally

1:22:41Step into ComputeVoxelIrradianceAt() and compare the Txy and Check

🏃

1:22:41Step into ComputeVoxelIrradianceAt() and compare the Txy and Check

🏃

1:22:41Step into ComputeVoxelIrradianceAt() and compare the Txy and Check

🏃

1:24:48Weld GetOctahedralOffset() into ComputeVoxelIrradianceAt() to facilitate closer comparison

1:27:47Hit our assertion in ComputeVoxelIrradianceAt(), and compare the check and newly computed values

🏃

1:27:47Hit our assertion in ComputeVoxelIrradianceAt(), and compare the check and newly computed values

🏃

1:27:47Hit our assertion in ComputeVoxelIrradianceAt(), and compare the check and newly computed values

🏃

1:29:49Weld OctahedralFromUnitVector() into ComputeVoxelIrradianceAt() to facilitate comparison

1:32:25Hit our assertion in ComputeVoxelIrradianceAt(), and compare the check and newly computed values

🏃

1:32:25Hit our assertion in ComputeVoxelIrradianceAt(), and compare the check and newly computed values

🏃

1:32:25Hit our assertion in ComputeVoxelIrradianceAt(), and compare the check and newly computed values

🏃

1:32:54Fix ComputeVoxelIrradianceAt() to use AbsoluteValue() when computing the OneNorm

1:33:26We run successfully

🏃

1:33:26We run successfully

🏃

1:33:26We run successfully

🏃

1:33:35Remove the checking code from ComputeVoxelIrradianceAt()

1:33:51Our ray casting performance is improving

🏃

1:33:51Our ray casting performance is improving

🏃

1:33:51Our ray casting performance is improving

🏃

1:35:38hhlightprof total seconds elapsed: 5.244874

🏃

1:35:38hhlightprof total seconds elapsed: 5.244874

🏃

1:35:38hhlightprof total seconds elapsed: 5.244874

🏃

1:37:33Inspect the assembly of ComputeVoxelIrradianceAt()

🏃

1:37:33Inspect the assembly of ComputeVoxelIrradianceAt()

🏃

1:37:33Inspect the assembly of ComputeVoxelIrradianceAt()

🏃

1:38:26Replace Extract0(), Extract1() and Extract2() with ConvertF32() and ConvertS32()

1:42:08Our lighting looks the same

🏃

1:42:08Our lighting looks the same

🏃

1:42:08Our lighting looks the same

🏃

1:42:15hhlightprof total seconds elapsed: 5.055217

🏃

1:42:15hhlightprof total seconds elapsed: 5.055217

🏃

1:42:15hhlightprof total seconds elapsed: 5.055217

🏃

1:43:11Change ComputeVoxelIrradianceAt() to return the f32_4x ResultRGB, for the callers to use directly

1:46:13Our lighting looks the same

🏃

1:46:13Our lighting looks the same

🏃

1:46:13Our lighting looks the same

🏃

1:46:18hhlightprof total seconds elapsed: 4.963887

🏃

1:46:18hhlightprof total seconds elapsed: 4.963887

🏃

1:46:18hhlightprof total seconds elapsed: 4.963887

🏃

1:47:14Weld ComputeVoxelIrradianceAt() straight in to GridRayCast(), to save computing values twice

1:49:16Our lighting looks the same

🏃

1:49:16Our lighting looks the same

🏃

1:49:16Our lighting looks the same

🏃

1:49:30hhlightprof total seconds elapsed: 4.701094

🏃

1:49:30hhlightprof total seconds elapsed: 4.701094

🏃

1:49:30hhlightprof total seconds elapsed: 4.701094

🏃

1:50:40Seek further improvements to GridRayCast()

📖

1:50:40Seek further improvements to GridRayCast()

📖

1:50:40Seek further improvements to GridRayCast()

📖

1:54:19Our lighting looks the same

🏃

1:54:19Our lighting looks the same

🏃

1:54:19Our lighting looks the same

🏃

1:54:22Q&A

🗩

1:54:22Q&A

🗩

1:54:22Q&A

🗩

1:54:43golido3868 Q: Sorry to be off-topic. I've finished all the five days in the Intro to C and it was awesome. But there's a huge gap between the intro and the main course that I'm not able to fully understand. What's your suggestion? I'm new to programming, started reading K&R recently.⁹

🗪

1:54:43golido3868 Q: Sorry to be off-topic. I've finished all the five days in the Intro to C and it was awesome. But there's a huge gap between the intro and the main course that I'm not able to fully understand. What's your suggestion? I'm new to programming, started reading K&R recently.⁹

🗪

1:54:43golido3868 Q: Sorry to be off-topic. I've finished all the five days in the Intro to C and it was awesome. But there's a huge gap between the intro and the main course that I'm not able to fully understand. What's your suggestion? I'm new to programming, started reading K&R recently.⁹

🗪

1:56:35Plug Star Code Galaxy¹⁰

📖

1:56:35Plug Star Code Galaxy¹⁰

📖

1:56:35Plug Star Code Galaxy¹⁰

📖

1:57:39Close it down

🗩

1:57:39Close it down

🗩

1:57:39Close it down

🗩

Keyboard Navigation

Global Keys

Menu toggling

In-Menu and Index Controls

Quotes and References Menus and Index

Quotes, References and Credits Menus

Filter Menu

Filter and Link Menus

Credits Menu