sean-m
/
kendryte-freertos-sdk
mirror of https://github.com/kendryte/kendryte-freertos-sdk.git
1
0
Fork 0
Code Issues Releases Wiki Activity

add kpu driver

fix/double
jiangxiangbing 2019-03-02 14:37:16 +08:00
parent 863d783ea0
commit 23b78c9d22
6 changed files with 1011 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

512
lib/bsp/device/kpu.cpp Normal file
View File

@ -0,0 +1,512 @@
/* Copyright 2018 Canaan Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <stdlib.h>
#include <string.h>
#include <FreeRTOS.h>
#include <dmac.h>
#include <hal.h>
#include <kernel/driver_impl.hpp>
#include <kpu.h>
#include <sysctl.h>
#include <math.h>
using namespace sys;
#define COMMON_ENTRY \
semaphore_lock locker(free_mutex_);
class k_model_context : public heap_object, public free_object_access
{
public:
k_model_context(uint8_t *buffer)
{
uintptr_t base_addr = (uintptr_t)buffer;
kpu_model_header_t *header = (kpu_model_header_t *)buffer;
model_buffer_ = buffer;
layer_headers_ = (kpu_model_layer_header_t *)(base_addr + sizeof(kpu_model_header_t));
layers_length_ = header->layers_length;
body_start_ = (uint8_t *)(base_addr + sizeof(kpu_model_header_t) + 8 * header->layers_length);
storage_ = std::make_unique<uint8_t[]>(header->main_mem_usage);
main_buffer_ = { storage_.get(), ptrdiff_t(header->main_mem_usage) };
}
virtual void on_first_open() override
{
}
virtual void on_last_close() override
{
}
void get(kpu_model_context_t *ctx)
{
ctx->body_start = body_start_;
ctx->model_buffer = model_buffer_;
ctx->main_buffer = main_buffer_.data();
ctx->layer_headers = layer_headers_;
ctx->layers_length = layers_length_;
}
private:
uint8_t *model_buffer_;
kpu_model_layer_header_t *layer_headers_;
uint8_t *body_start_;
uint32_t layers_length_;
gsl::span<uint8_t> main_buffer_;
std::unique_ptr<uint8_t[]> storage_;
};
class k_kpu_driver : public kpu_driver, public static_object, public free_object_access
{
public:
k_kpu_driver(uintptr_t base_addr, sysctl_clock_t clock, sysctl_dma_select_t dma_req)
: kpu_(*reinterpret_cast<volatile kpu_config_t *>(base_addr)), clock_(clock), dma_req_(dma_req)
{
completion_event_ = xSemaphoreCreateBinary();
}
virtual void install() override
{
free_mutex_ = xSemaphoreCreateMutex();
sysctl_clock_disable(clock_);
}
virtual void on_first_open() override
{
sysctl_clock_enable(clock_);
}
virtual void on_last_close() override
{
sysctl_clock_disable(clock_);
}
virtual handle_t model_load_from_buffer(uint8_t *buffer) override
{
return system_alloc_handle(make_accessor(make_object<k_model_context>(buffer)));
}
virtual int run(handle_t context, const uint8_t *src, uint8_t **output, size_t *output_size) override
{
COMMON_ENTRY;
auto model_context = system_handle_to_object(context).as<k_model_context>();
model_context->get(&ctx_);
dma_ch_ = dma_open_free();
ctx_.current_layer = 0;
ctx_.current_body = ctx_.body_start;
kpu_model_header_t *header = (kpu_model_header_t *)ctx_.model_buffer;
kpu_.interrupt_clear.reg = 7;
kpu_.fifo_threshold.data.fifo_full_threshold = 10;
kpu_.fifo_threshold.data.fifo_empty_threshold = 1;
kpu_.fifo_threshold.data.reserved = 0;
kpu_.eight_bit_mode.data.eight_bit_mode = header->flags & 1;
kpu_.eight_bit_mode.data.reserved = 0;
kpu_.interrupt_mask.data.calc_done_int = 1;
kpu_.interrupt_mask.data.layer_cfg_almost_empty_int = 0;
kpu_.interrupt_mask.data.layer_cfg_almost_full_int = 1;
kpu_.interrupt_mask.data.reserved = 0;
pic_set_irq_priority(IRQN_AI_INTERRUPT, 1);
pic_set_irq_handler(IRQN_AI_INTERRUPT, kpu_isr_handle, this);
pic_set_irq_enable(IRQN_AI_INTERRUPT, 1);
kpu_model_layer_header_t *first_layer_header = ctx_.layer_headers;
if (first_layer_header->type != KL_K210_CONV)
return -1;
kpu_model_conv_layer_argument_t *first_layer = (kpu_model_conv_layer_argument_t *)ctx_.body_start;
kpu_layer_argument_t *layer_arg = (kpu_layer_argument_t *)(ctx_.model_buffer + first_layer->layer_offset);
if ((layer_arg->image_size.data.i_row_wid + 1) % 64 != 0)
{
kpu_input_with_padding(layer_arg, src);
ai_step_not_isr();
}
else
{
kpu_input_dma(layer_arg, src);
}
while (!done_flag_)
{
if(xSemaphoreTake(completion_event_, portMAX_DELAY) == pdTRUE)
{
if (ctx_.current_layer != ctx_.layers_length)
{
while(ai_step() == 1)
;
}
else
{
kpu_done();
}
}
}
*output = output_address_;
*output_size = output_size_;
return 0;
}
private:
static void kpu_isr_handle(void *userdata)
{
auto &driver = *reinterpret_cast<k_kpu_driver *>(userdata);
driver.kpu_.interrupt_clear.data.calc_done_int = 1;
driver.kpu_.interrupt_clear.data.layer_cfg_almost_empty_int = 1;
driver.kpu_.interrupt_clear.data.layer_cfg_almost_full_int = 1;
driver.kpu_.interrupt_clear.data.reserved = 0;
driver.kpu_.interrupt_mask.data.calc_done_int = 1;
driver.kpu_.interrupt_mask.data.layer_cfg_almost_empty_int = 1;
driver.kpu_.interrupt_mask.data.layer_cfg_almost_full_int = 1;
driver.kpu_.interrupt_mask.data.reserved = 0;
BaseType_t xHigherPriorityTaskWoken = pdFALSE;
xSemaphoreGiveFromISR(driver.completion_event_, &xHigherPriorityTaskWoken);
if (xHigherPriorityTaskWoken)
{
portYIELD_FROM_ISR();
}
}
int ai_step()
{
uint32_t cnt_layer_id = ctx_.current_layer++;
uint8_t *layer_body = ctx_.current_body;
kpu_model_layer_header_t *cnt_layer_header = ctx_.layer_headers + cnt_layer_id;
ctx_.current_body += cnt_layer_header->body_size;
switch (cnt_layer_header->type)
{
case KL_GLOBAL_AVERAGE_POOL2D:
kpu_global_average_pool2d((kpu_model_gap2d_layer_argument_t *)layer_body);
break;
case KL_QUANTIZE:
kpu_quantize((kpu_model_quantize_layer_argument_t *)layer_body);
break;
case KL_DEQUANTIZE:
kpu_dequantize((kpu_model_dequantize_layer_argument_t *)layer_body);
break;
case KL_L2_NORMALIZATION:
kpu_l2_normalization((kpu_model_l2_norm_layer_argument_t *)layer_body);
break;
case KL_K210_CONV:
kpu_conv((kpu_model_conv_layer_argument_t *)layer_body);
return 0;
case KL_K210_ADD_PADDING:
kpu_add_padding((kpu_model_add_padding_layer_argument_t *)layer_body);
break;
case KL_K210_REMOVE_PADDING:
kpu_remove_padding((kpu_model_remove_padding_layer_argument_t *)layer_body);
break;
default:
configASSERT("Layer is not supported.");
}
if (cnt_layer_id != (ctx_.layers_length - 1))
{
return 1;
}
else
{
kpu_done();
return 0;
}
}
void kpu_input_with_padding(kpu_layer_argument_t *layer, const uint8_t *src)
{
size_t width = layer->image_size.data.i_row_wid + 1;
size_t height = layer->image_size.data.i_col_high + 1;
size_t channels = layer->image_channel_num.data.i_ch_num + 1;
uint8_t *dest = (uint8_t *)(uintptr_t)(AI_IO_BASE_ADDR + layer->image_addr.data.image_src_addr * 64);
size_t oc, y, x;
uint32_t row_padding;
uint32_t row_group;
uint32_t row_length;
if (width <= 16)
{
row_padding = 16;
row_group = 4;
row_length = 1;
}
else if (width <= 32)
{
row_padding = 32;
row_group = 2;
row_length = 1;
}
else
{
row_padding = 64;
row_group = 1;
row_length = (width + 63) / 64;
}
for (oc = 0; oc < channels; oc++)
{
uint8_t *channel_origin = dest + oc / row_group * row_length * height * 64 + oc % row_group * row_padding;
for (y = 0; y < height; y++)
{
uint8_t *y_origin = channel_origin + y * row_length * 64;
for (x = 0; x < width; x++)
y_origin[x] = *src++;
}
}
}
void ai_step_not_isr()
{
portENTER_CRITICAL();
ai_step();
vPortExitCritical();
}
void kpu_input_dma(kpu_layer_argument_t *layer, const uint8_t *src)
{
uint64_t input_size = layer->kernel_calc_type_cfg.data.channel_switch_addr * 64 * (layer->image_channel_num.data.i_ch_num + 1);
dma_set_request_source(dma_ch_, dma_req_);
dma_transmit_async(dma_ch_, src, (void *)(uintptr_t)((uint8_t *)AI_IO_BASE_ADDR + layer->image_addr.data.image_src_addr * 64), 1, 1, sizeof(uint64_t), input_size / 8, 16, completion_event_);
}
int kpu_done()
{
kpu_.interrupt_clear.data.calc_done_int = 1;
kpu_.interrupt_clear.data.layer_cfg_almost_empty_int = 1;
kpu_.interrupt_clear.data.layer_cfg_almost_full_int = 1;
kpu_.interrupt_clear.data.reserved = 0;
kpu_.interrupt_mask.data.calc_done_int = 1;
kpu_.interrupt_mask.data.layer_cfg_almost_empty_int = 1;
kpu_.interrupt_mask.data.layer_cfg_almost_full_int = 1;
kpu_.interrupt_mask.data.reserved = 0;
kpu_model_header_t *header = (kpu_model_header_t *)ctx_.model_buffer;
output_address_ = ctx_.main_buffer + header->output_address;
output_size_ = header->output_size;
done_flag_ = 1;
return 0;
}
void kpu_conv(kpu_model_conv_layer_argument_t *arg)
{
kpu_layer_argument_t *layer = (kpu_layer_argument_t *)(ctx_.model_buffer + arg->layer_offset);
layer->kernel_load_cfg.data.para_start_addr = (uintptr_t)(ctx_.model_buffer + arg->weights_offset);
layer->kernel_pool_type_cfg.data.bwsx_base_addr = (uintptr_t)(ctx_.model_buffer + arg->bn_offset);
layer->kernel_calc_type_cfg.data.active_addr = (uintptr_t)(ctx_.model_buffer + arg->act_offset);
if (arg->flags & KLF_MAIN_MEM_OUT)
{
uint8_t *dest = ctx_.main_buffer + arg->main_mem_out_address;
layer->dma_parameter.data.send_data_out = 1;
dma_set_request_source(dma_ch_, dma_req_);
dma_transmit_async(dma_ch_, (void *)(&kpu_.fifo_data_out), dest, 0, 1, sizeof(uint64_t), (layer->dma_parameter.data.dma_total_byte + 8) / 8, 8, completion_event_);
kpu_send_layer(layer);
}
else
{
kpu_send_layer(layer);
kpu_.interrupt_mask.data.calc_done_int = 1;
kpu_.interrupt_mask.data.layer_cfg_almost_empty_int = 0;
kpu_.interrupt_mask.data.layer_cfg_almost_full_int = 1;
kpu_.interrupt_mask.data.reserved = 0;
}
}
void kpu_send_layer(const kpu_layer_argument_t *layer)
{
kpu_.layer_argument_fifo = layer->interrupt_enabe.reg;
kpu_.layer_argument_fifo = layer->image_addr.reg;
kpu_.layer_argument_fifo = layer->image_channel_num.reg;
kpu_.layer_argument_fifo = layer->image_size.reg;
kpu_.layer_argument_fifo = layer->kernel_pool_type_cfg.reg;
kpu_.layer_argument_fifo = layer->kernel_load_cfg.reg;
kpu_.layer_argument_fifo = layer->kernel_offset.reg;
kpu_.layer_argument_fifo = layer->kernel_calc_type_cfg.reg;
kpu_.layer_argument_fifo = layer->write_back_cfg.reg;
kpu_.layer_argument_fifo = layer->conv_value.reg;
kpu_.layer_argument_fifo = layer->conv_value2.reg;
kpu_.layer_argument_fifo = layer->dma_parameter.reg;
}
void kpu_global_average_pool2d(kpu_model_gap2d_layer_argument_t *arg)
{
const float *src = (const float *)(ctx_.main_buffer + arg->main_mem_in_address);
float *dest = (float *)(ctx_.main_buffer + arg->main_mem_out_address);
size_t oc, channels = arg->channels, kernel_size = arg->kernel_size;
for (oc = 0; oc < channels; oc++)
{
float sum = 0.f;
size_t i;
for (i = 0; i < kernel_size; i++)
sum += *src++;
dest[oc] = sum / kernel_size;
}
}
void kpu_quantize(kpu_model_quantize_layer_argument_t *arg)
{
size_t width = arg->width;
size_t height = arg->height;
size_t channels = arg->channels;
const float *src = (const float *)(ctx_.main_buffer + arg->main_mem_in_address);;
const kpu_model_quant_param_t q = arg->quant_param;
if (arg->flags & KLF_MAIN_MEM_OUT)
{
uint8_t *dest = (uint8_t *)(ctx_.main_buffer + arg->mem_out_address);
size_t i, count = width * height * channels;
for (i = 0; i < count; i++)
{
int value = (*src++ - q.bias) / q.scale;
if (value < 0) value = 0;
if (value > 0xFF) value = 0xFF;
*dest++ = (uint8_t)value;
}
}
else
{
uint8_t *dest = (uint8_t *)AI_IO_BASE_ADDR + arg->mem_out_address * 64;
size_t oc, y, x;
uint32_t row_padding;
uint32_t row_group;
uint32_t row_length;
if (width <= 16)
{
row_padding = 16;
row_group = 4;
row_length = 1;
}
else if (width <= 32)
{
row_padding = 32;
row_group = 2;
row_length = 1;
}
else
{
row_padding = 64;
row_group = 1;
row_length = (width + 63) / 64;
}
for (oc = 0; oc < channels; oc++)
{
uint8_t *channel_origin = dest + oc / row_group * row_length * height * 64 + oc % row_group * row_padding;
for (y = 0; y < height; y++)
{
uint8_t *y_origin = channel_origin + y * row_length * 64;
for (x = 0; x < width; x++)
{
int value = (*src++ - q.bias) / q.scale;
if (value < 0) value = 0;
if (value > 0xFF) value = 0xFF;
y_origin[x] = (uint8_t)value;
}
}
}
}
}
void kpu_dequantize(kpu_model_dequantize_layer_argument_t *arg)
{
const uint8_t *src = (const uint8_t *)(ctx_.main_buffer + arg->main_mem_in_address);
float *dest = (float *)(ctx_.main_buffer + arg->main_mem_out_address);
size_t oc, count = arg->count;
const kpu_model_quant_param_t q = arg->quant_param;
for (oc = 0; oc < count; oc++)
dest[oc] = *src++ * q.scale + q.bias;
}
void kpu_l2_normalization(kpu_model_l2_norm_layer_argument_t *arg)
{
const float *src = (const float *)(ctx_.main_buffer + arg->main_mem_in_address);
float *dest = (float *)(ctx_.main_buffer + arg->main_mem_out_address);
size_t oc, channels = arg->channels;
float sum = 0.f;
const float epsilon = 1e-10f;
for (oc = 0; oc < channels; oc++)
sum += src[oc] * src[oc];
if (sum < epsilon)
sum = epsilon;
sum = 1.f / sqrtf(sum);
for (oc = 0; oc < channels; oc++)
dest[oc] = src[oc] * sum;
}
void kpu_add_padding(kpu_model_add_padding_layer_argument_t *arg)
{
const uint8_t *src = (const uint8_t *)(ctx_.main_buffer + arg->main_mem_in_address);
uint8_t *dest = (uint8_t *)AI_IO_BASE_ADDR + arg->kpu_mem_out_address * 64L;
uint32_t row_padding = 16;
uint32_t row_group = 4;
uint32_t row_length = 1;
uint32_t height = 4;
uint32_t oc, x, y, channels = arg->channels;
for (oc = 0; oc < channels; oc++)
{
uint8_t *channel_origin = dest + oc / row_group * row_length * height * 64 + oc % row_group * row_padding;
for (y = 0; y < 1; y++)
{
uint8_t *y_origin = channel_origin + y * row_length * 64;
for (x = 0; x < 1; x++)
y_origin[x] = *src++;
}
}
}
void kpu_remove_padding(kpu_model_remove_padding_layer_argument_t *arg)
{
const uint8_t *src = (const uint8_t *)(ctx_.main_buffer + arg->main_mem_in_address);
uint8_t *dest = (uint8_t *)(ctx_.main_buffer + arg->main_mem_out_address);
uint32_t oc, channels = arg->channels;
for (oc = 0; oc < channels; oc++)
*dest++ = src[oc * 16];
}
private:
volatile kpu_config_t &kpu_;
sysctl_clock_t clock_;
sysctl_dma_select_t dma_req_;
SemaphoreHandle_t free_mutex_;
uintptr_t dma_ch_;
SemaphoreHandle_t completion_event_;
uint8_t done_flag_ = 0;
uint8_t *output_address_;
size_t output_size_;
kpu_model_context_t ctx_;
};
static k_kpu_driver dev0_driver(AI_BASE_ADDR, SYSCTL_CLOCK_AI, SYSCTL_DMA_SELECT_AI_RX_REQ);
driver &g_kpu_driver_kpu0 = dev0_driver;

3
lib/bsp/device/registry.cpp
View File

@ -70,6 +70,8 @@ extern driver &g_wdt_driver_wdt1;
extern driver &g_rtc_driver_rtc0;
extern driver &g_kpu_driver_kpu0;
driver_registry_t sys::g_system_drivers[] = {
{ "/dev/uart1", { std::in_place, &g_uart_driver_uart0 } },
{ "/dev/uart2", { std::in_place, &g_uart_driver_uart1 } },
@ -122,6 +124,7 @@ driver_registry_t sys::g_system_drivers[] = {
{ "/dev/wdt1", { std::in_place, &g_wdt_driver_wdt1 } },
{ "/dev/rtc0", { std::in_place, &g_rtc_driver_rtc0 } },
{ "/dev/kpu0", { std::in_place, &g_kpu_driver_kpu0 } },
{}
};

25
lib/freertos/include/devices.h
View File

@ -832,6 +832,31 @@ void rtc_get_datetime(handle_t file, struct tm *datetime);
*/
void rtc_set_datetime(handle_t file, const struct tm *datetime);
/**
* @brief Load model from buffer
*
* @param[in] buffer model data
*
* @return result
* - 0 Fail
* - other The kpu context handle
*/
handle_t kpu_model_load_from_buffer(uint8_t *buffer);
/**
* @brief KPU run and get output data.
*
* @param[in] context The kpu context handle
* @param[in] src The src data
* @param[in] output The address of the kpu output data address.
* @param[in] output_size The address of output data size
*
* @return result
* - 0 Success
* - other Fail
*/
int kpu_run(handle_t context, const uint8_t *src, uint8_t **output, size_t *output_size);
#ifdef __cplusplus
}
#endif

7
lib/freertos/include/kernel/driver.hpp
View File

@ -352,6 +352,13 @@ public:
virtual void set_datetime(const struct tm &datetime) = 0;
};
class kpu_driver : public driver
{
public:
virtual handle_t model_load_from_buffer(uint8_t *buffer) = 0;
virtual int run(handle_t context, const uint8_t *src, uint8_t **output, size_t *output_size) = 0;
};
class custom_driver : public driver
{
public:

15
lib/freertos/kernel/devices.cpp
View File

@ -64,6 +64,7 @@ static const char dummy_driver_name[] = "";
uintptr_t fft_file_;
uintptr_t aes_file_;
uintptr_t sha256_file_;
uintptr_t kpu_file_;
extern UBaseType_t uxCPUClockRate;
@ -125,6 +126,7 @@ void install_drivers()
fft_file_ = io_open("/dev/fft0");
aes_file_ = io_open("/dev/aes0");
sha256_file_ = io_open("/dev/sha256");
kpu_file_ = io_open("/dev/kpu0");
}
static _file *io_alloc_file(object_accessor<object_access> object)
@ -790,6 +792,19 @@ void rtc_set_datetime(handle_t file, const struct tm *datetime)
rtc->set_datetime(*datetime);
}
/* KPU */
handle_t kpu_model_load_from_buffer(uint8_t *buffer)
{
COMMON_ENTRY_FILE(kpu_file_, kpu);
return kpu->model_load_from_buffer(buffer);
}
int kpu_run(handle_t context, const uint8_t *src, uint8_t **output, size_t *output_size)
{
COMMON_ENTRY_FILE(kpu_file_, kpu);
return kpu->run(context, src, output, output_size);
}
/* HAL */
static uintptr_t pic_file_;

449
lib/hal/include/kpu.h Normal file
View File

@ -0,0 +1,449 @@
/* Copyright 2018 Canaan Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef _KPU_H
#define _KPU_H
#include <stdint.h>
#include <stddef.h>
#include <platform.h>
#ifdef __cplusplus
extern "C" {
#endif
#define ALIGN_UP(x, align) ((x + (align - 1)) & (~(align - 1)))
#define KMODEL_HEADER_SIZE_V2(layers) (ALIGN_UP(sizeof(kpu_model_header_t) + sizeof(kpu_model_layer_metadata_v2_t) * layers, 8) + sizeof(kpu_layer_argument_t) * layers)
typedef struct
{
union
{
uint64_t reg;
struct
{
uint64_t int_en:1;
uint64_t ram_flag:1;
uint64_t full_add:1;
uint64_t depth_wise_layer:1;
uint64_t reserved:60;
} data;
} interrupt_enabe;
union
{
uint64_t reg;
struct
{
uint64_t image_src_addr:15;
uint64_t reserved0:17;
uint64_t image_dst_addr:15;
uint64_t reserved1:17;
} data;
} image_addr;
union
{
uint64_t reg;
struct
{
uint64_t i_ch_num:10;
uint64_t reserved0:22;
uint64_t o_ch_num:10;
uint64_t reserved1:6;
uint64_t o_ch_num_coef:10;
uint64_t reserved2:6;
} data;
} image_channel_num;
union
{
uint64_t reg;
struct
{
uint64_t i_row_wid:10;
uint64_t i_col_high:9;
uint64_t reserved0:13;
uint64_t o_row_wid:10;
uint64_t o_col_high:9;
uint64_t reserved1:13;
} data;
} image_size;
union
{
uint64_t reg;
struct
{
uint64_t kernel_type:3;
uint64_t pad_type:1;
uint64_t pool_type:4;
uint64_t first_stride:1;
uint64_t bypass_conv:1;
uint64_t load_para:1;
uint64_t reserved0:5;
uint64_t dma_burst_size:8;
uint64_t pad_value:8;
uint64_t bwsx_base_addr:32;
} data;
} kernel_pool_type_cfg;
union
{
uint64_t reg;
struct
{
uint64_t load_coor:1;
uint64_t load_time:6;
uint64_t reserved0:8;
uint64_t para_size:17;
uint64_t para_start_addr:32;
} data;
} kernel_load_cfg;
union
{
uint64_t reg;
struct
{
uint64_t coef_column_offset:4;
uint64_t coef_row_offset:12;
uint64_t reserved0:48;
} data;
} kernel_offset;
union
{
uint64_t reg;
struct
{
uint64_t channel_switch_addr:15;
uint64_t reserved:1;
uint64_t row_switch_addr:4;
uint64_t coef_size:8;
uint64_t coef_group:3;
uint64_t load_act:1;
uint64_t active_addr:32;
} data;
} kernel_calc_type_cfg;
union
{
uint64_t reg;
struct
{
uint64_t wb_channel_switch_addr:15;
uint64_t reserved0:1;
uint64_t wb_row_switch_addr:4;
uint64_t wb_group:3;
uint64_t reserved1:41;
} data;
} write_back_cfg;
union
{
uint64_t reg;
struct
{
uint64_t shr_w:4;
uint64_t shr_x:4;
uint64_t arg_w:24;
uint64_t arg_x:24;
uint64_t reserved0:8;
} data;
} conv_value;
union
{
uint64_t reg;
struct
{
uint64_t arg_add:40;
uint64_t reserved:24;
} data;
} conv_value2;
union
{
uint64_t reg;
struct
{
uint64_t send_data_out:1;
uint64_t reserved:15;
uint64_t channel_byte_num:16;
uint64_t dma_total_byte:32;
} data;
} dma_parameter;
} kpu_layer_argument_t;
typedef struct
{
union
{
uint64_t reg;
struct
{
uint64_t shift_number:8;
uint64_t y_mul:16;
uint64_t x_start:36;
} data;
} activate_para[16];
union
{
uint64_t reg;
struct
{
uint8_t result_bias[8];
} data;
} activate_para_bias0;
union
{
uint64_t reg;
struct
{
uint8_t result_bias[8];
} data;
} activate_para_bias1;
} kpu_activate_table_t;
typedef struct
{
union
{
uint64_t reg;
struct
{
uint64_t norm_mul:24;
uint64_t norm_add:32;
uint64_t norm_shift:4;
} data;
} batchnorm;
} kpu_batchnorm_argument_t;
typedef struct
{
union
{
uint64_t reg;
struct
{
uint16_t weight[9];
} data;
} weights;
} kpu_weights_kernel_16_3x3_t;
typedef struct
{
uint64_t calc_done_int:1;
uint64_t layer_cfg_almost_empty_int:1;
uint64_t layer_cfg_almost_full_int:1;
uint64_t reserved:61;
} kpu_config_interrupt_t;
typedef struct
{
uint64_t fifo_full_threshold:4;
uint64_t fifo_empty_threshold:4;
uint64_t reserved:56;
} kpu_config_fifo_threshold_t;
typedef struct
{
uint64_t dma_fifo_flush_n:1;
uint64_t gs_fifo_flush_n:1;
uint64_t cfg_fifo_flush_n:1;
uint64_t cmd_fifo_flush_n:1;
uint64_t resp_fifo_flush_n:1;
uint64_t reserved:59;
} kpu_config_fifo_ctrl_t;
typedef struct
{
uint64_t eight_bit_mode:1;
uint64_t reserved:63;
} kpu_config_eight_bit_mode_t;
typedef struct
{
volatile uint64_t layer_argument_fifo;
volatile union
{
uint64_t reg;
kpu_config_interrupt_t data;
} interrupt_status;
volatile union
{
uint64_t reg;
kpu_config_interrupt_t data;
} interrupt_raw;
volatile union {
uint64_t reg;
kpu_config_interrupt_t data;
} interrupt_mask;
volatile union
{
uint64_t reg;
kpu_config_interrupt_t data;
} interrupt_clear;
volatile union
{
uint64_t reg;
kpu_config_fifo_threshold_t data;
} fifo_threshold;
volatile uint64_t fifo_data_out;
volatile union
{
uint64_t reg;
kpu_config_fifo_ctrl_t data;
} fifo_ctrl;
volatile union
{
uint64_t reg;
kpu_config_eight_bit_mode_t data;
} eight_bit_mode;
} kpu_config_t;
typedef struct
{
uint32_t version;
uint32_t flags;
uint32_t layers_length;
uint32_t max_start_address;
uint32_t main_mem_usage;
uint32_t output_address;
uint32_t output_size;
} kpu_model_header_t;
typedef enum
{
KL_GLOBAL_AVERAGE_POOL2D = 0,
KL_QUANTIZE = 1,
KL_DEQUANTIZE = 2,
KL_L2_NORMALIZATION = 3,
KL_K210_CONV = 4,
KL_K210_ADD_PADDING = 5,
KL_K210_REMOVE_PADDING = 6,
_KL_MAX_COUNT
} kpu_model_layer_type_t;
typedef struct
{
uint32_t type;
uint32_t body_size;
} kpu_model_layer_header_t;
typedef enum
{
KLF_NONE = 0,
KLF_MAIN_MEM_OUT = 1
} kpu_model_layer_flags_t;
typedef struct
{
float scale;
float bias;
} kpu_model_quant_param_t;
typedef struct
{
uint32_t flags;
uint32_t main_mem_out_address;
uint32_t layer_offset;
uint32_t weights_offset;
uint32_t bn_offset;
uint32_t act_offset;
} kpu_model_conv_layer_argument_t;
typedef struct
{
uint32_t flags;
uint32_t main_mem_in_address;
uint32_t main_mem_out_address;
uint32_t kernel_size;
uint32_t channels;
} kpu_model_gap2d_layer_argument_t;
typedef struct
{
uint32_t flags;
uint32_t main_mem_in_address;
uint32_t mem_out_address;
uint32_t width;
uint32_t height;
uint32_t channels;
kpu_model_quant_param_t quant_param;
} kpu_model_quantize_layer_argument_t;
typedef struct
{
uint32_t flags;
uint32_t main_mem_in_address;
uint32_t main_mem_out_address;
uint32_t count;
kpu_model_quant_param_t quant_param;
} kpu_model_dequantize_layer_argument_t;
typedef struct
{
uint32_t flags;
uint32_t main_mem_in_address;
uint32_t kpu_mem_out_address;
uint32_t channels;
} kpu_model_add_padding_layer_argument_t;
typedef struct
{
uint32_t flags;
uint32_t main_mem_in_address;
uint32_t main_mem_out_address;
uint32_t channels;
} kpu_model_remove_padding_layer_argument_t;
typedef struct
{
uint32_t flags;
uint32_t main_mem_in_address;
uint32_t main_mem_out_address;
uint32_t channels;
} kpu_model_l2_norm_layer_argument_t;
typedef struct
{
uint8_t *model_buffer;
uint8_t *main_buffer;
kpu_model_layer_header_t *layer_headers;
uint8_t *body_start;
uint32_t layers_length;
volatile uint32_t current_layer;
uint8_t * volatile current_body;
} kpu_model_context_t;
#ifdef __cplusplus
}
#endif
#endif