1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
|
#pragma once
#include "rend_internal.h"
#include "rend_vk_internal.h"
#include <vulkan/vulkan_core.h>
/* NOTE(vasco):
* Check out https://kylehalladay.com/blog/tutorial/2017/12/13/Custom-Allocators-Vulkan.html
* for a basic grasp on what it entails to actually write a memory allocator for vulkan.
*
* This allocator isnt anything particularly amazing but it works so I'm keeping this code.
* Based on the framework presented here we could implement other more efficient allocators.
* Which is what I did for the arena allocator.
*/
#if 0
typedef struct RendVkAddress {
VkDeviceMemory handle;
uint32_t type;
uint32_t id;
VkDeviceSize size;
VkDeviceSize offset;
} RendVkAddress;
typedef struct RendVkPoolLayout {
uint64_t offset, size;
} RendVkPoolLayout;
typedef struct RendVKPoolBlock {
RendVkAddress address;
RendVkPoolLayout *layout_darr;
uint8_t reserved;
} RendVKPoolBlock;
typedef struct RendVkPoolMemory {
RendVKPoolBlock *block_darr;
} RendVkPoolMemory;
struct RendVkPoolAllocator {
VkAllocationCallbacks *allocator;
size_t *mem_type_alloc_sizes; // allocation size per type of memory available on the gpu
RendVkPoolMemory *mem_pools; // memory pools per type of memory available on the gpu
VkDevice logical_device; // virtual device
VkPhysicalDevice physical_device; // physical device we are allocating memory from
VkDeviceSize page_size; // allocations must respect the physical limitations of the gpu
VkDeviceSize block_min_size; // minimum size per block of memory
uint64_t total_allocations;
uint32_t memory_type_count;
};
static RendVkPoolAllocator
rend_vk_pool_create(VkDevice logical_device, VkPhysicalDevice physical_device, VkPhysicalDeviceLimits device_limits, VkAllocationCallbacks *allocator)
{
RendVkPoolAllocator pool = {
.allocator = allocator,
.mem_type_alloc_sizes = NULL,
.mem_pools = NULL,
.logical_device = logical_device,
.physical_device = physical_device,
.page_size = 0,
.block_min_size = 0,
.total_allocations = 0,
.memory_type_count = 0,
};
VkPhysicalDeviceMemoryProperties mem_properties;
vkGetPhysicalDeviceMemoryProperties(physical_device, &mem_properties);
pool.mem_type_alloc_sizes = malloc(mem_properties.memoryTypeCount * sizeof *pool.mem_type_alloc_sizes);
memset(pool.mem_type_alloc_sizes, 0, mem_properties.memoryTypeCount * sizeof *pool.mem_type_alloc_sizes);
pool.mem_pools = malloc(mem_properties.memoryTypeCount * sizeof *pool.mem_pools);
memset(pool.mem_pools, 0, mem_properties.memoryTypeCount * sizeof *pool.mem_pools);
pool.memory_type_count = mem_properties.memoryTypeCount;
pool.page_size = device_limits.bufferImageGranularity;
pool.block_min_size = pool.page_size * 10;
return pool;
}
static uint32_t
rend_vk_pool_add_block(RendVkPoolAllocator *pool, VkDeviceSize size, uint32_t memory_type, VkMemoryPropertyFlags properties, bool fit_to_alloc)
{
VkDeviceSize new_pool_size = size * 2;
new_pool_size = (new_pool_size < pool->block_min_size) ? pool->block_min_size : new_pool_size;
VkMemoryAllocateInfo alloc_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO,
.allocationSize = new_pool_size,
.memoryTypeIndex = rend_vk_memory_find_index(pool->physical_device, memory_type, properties)
};
RendVKPoolBlock block = {0};
VkResult res = vkAllocateMemory(pool->logical_device, &alloc_info, pool->allocator, &block.address.handle);
block.address.type = memory_type;
block.address.size = new_pool_size;
RendVkPoolMemory *mem_pool = &pool->mem_pools[memory_type];
p_darray_push(mem_pool->block_darr, block);
RendVkPoolLayout layout = {
.offset = 0,
.size = new_pool_size
};
p_darray_push(mem_pool->block_darr[pool_size - 1].layout_darr, layout);
pool->total_allocations++;
size_t pool_size = p_darray_len(mem_pool->block_darr);
return pool_size - 1;
}
static RendVkMemory
rend_vk_pool_alloc(RendVkPoolAllocator *pool, VkDeviceSize size, uint32_t usage, uint32_t memory_type, VkMemoryPropertyFlags properties)
{
RendVkPoolMemory *mem_pool = &pool->mem_pools[memory_type];
VkDeviceSize requested_alloc_size = ((size / pool->page_size) + 1) * pool->page_size;
pool->mem_type_alloc_sizes[memory_type] += requested_alloc_size;
/* find free chunk for allocation
* TODO: free list?
*/
int whole_page = usage != VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT;
uint64_t block_idx = UINT64_MAX;
uint64_t span_idx = UINT64_MAX;
for (uint32_t i = 0; i < p_darray_len(mem_pool->block_darr); ++i) {
for (uint32_t j = 0; j < p_darray_len(mem_pool->block_darr[i].layout_darr); ++j) {
int valid = (whole_page) ? mem_pool->block_darr[i].layout_darr[j].offset == 0 : 1;
if (mem_pool->block_darr[i].layout_darr[j].size >= size && valid) {
block_idx = i;
span_idx = j;
}
}
}
if (block_idx == UINT64_MAX || span_idx == UINT64_MAX) {
block_idx = rend_vk_pool_add_block(pool, size, memory_type, properites, whole_page);
span_idx = 0;
}
mem_pool->block_darr[block_idx].reserved = whole_page;
RendVkMemory memory = {
.device_memory = mem_pool->block_darr[block_idx].address.handle,
.size = size,
.offset = mem_pool->block_darr[block_idx].layout_darr[span_idx].offset,
.type = memory_type,
.logical_device = pool->logical_device,
.allocator = pool->allocator,
.host_mapped_memory = 0,
.physical_device = pool->physical_device,
.properties = properties,
.id = block_idx,
};
/* mark chunck */
mem_pool->block_darr[block_idx].layout_darr[span_idx].offset += size;
mem_pool->block_darr[block_idx].layout_darr[span_idx].size -= size;
return memory;
}
static void
rend_vk_pool_free(RendVkPoolAllocator *pool, RendVkMemory *memory)
{
VkDeviceSize requested = ((memory->size / pool->page_size) + 1) * pool->page_size;
RendVkPoolMemory *mem_pool = &pool->mem_pools[memory->type];
pool->blocks_[allocation.id].pageReserved = false;
mem_pool->block_darr[block_idx].layout_darr[span_idx].offset += size;
mem_pool->block_darr[block_idx].layout_darr[span_idx].size -= size;
OffsetSize span = {allocation.offset, requestedAllocSize };
bool found = false;
uint32_t numLayoutMems = pool.blocks[allocation.id].layout.size();
for (uint32_t j = 0; j < numLayoutMems; ++j)
{
if (pool.blocks[allocation.id].layout[j].offset == requestedAllocSize +allocation.offset)
{
pool.blocks[allocation.id].layout[j].offset = allocation.offset;
pool.blocks[allocation.id].layout[j].size += requestedAllocSize;
found = true;
break;
}
}
if (!found)
{
state.memPools[allocation.type].blocks[allocation.id].layout.push_back(span);
state.memTypeAllocSizes[allocation.type] -= requestedAllocSize;
}
}
static void
rend_vk_pool_destroy(RendVkPoolAllocator *pool)
{
if (pool->mem_pools) {
for (size_t u = 0; u < pool->memory_type_count; ++u) {
RendVkPoolMemory mem_pool = pool->mem_pools[u];
for (size_t b = 0; b < p_darray_len(mem_pool.block_darr); ++b) {
RendVKPoolBlock block = mem_pool.block_darr[b];
p_darray_destroy(block.layout_darr);
}
p_darray_destroy(mem_pool.block_darr);
}
free(pool->mem_pools);
pool->mem_pools = 0;
}
if (pool->mem_type_alloc_sizes) {
free(pool->mem_type_alloc_sizes);
pool->mem_type_alloc_sizes = 0;
}
memset(pool, 0, sizeof *pool);
}
#endif
|
