Visible to Intel only — GUID: qcm1488476882488
Ixiasoft
1. Introduction
2. Avalon® -ST Multi-Channel Shared Memory FIFO Core
3. Avalon® -ST Single-Clock and Dual-Clock FIFO Cores
4. Avalon® -ST Serial Peripheral Interface Core
5. SPI Core
6. SPI Agent/JTAG to Avalon® Host Bridge Cores
7. Intel eSPI Agent Core
8. eSPI to LPC Bridge Core
9. Ethernet MDIO Core
10. Intel FPGA 16550 Compatible UART Core
11. UART Core
12. JTAG UART Core
13. Intel FPGA Avalon® Mailbox Core
14. Intel FPGA Avalon® Mutex Core
15. Intel FPGA Avalon® I2C (Host) Core
16. Intel FPGA I2C Agent to Avalon® -MM Host Bridge Core
17. Intel FPGA Avalon® Compact Flash Core
18. EPCS/EPCQA Serial Flash Controller Core
19. Intel FPGA Serial Flash Controller Core
20. Intel FPGA Serial Flash Controller II Core
21. Intel FPGA Generic QUAD SPI Controller Core
22. Intel FPGA Generic QUAD SPI Controller II Core
23. Interval Timer Core
24. Intel FPGA Avalon FIFO Memory Core
25. On-Chip Memory (RAM and ROM) Intel FPGA IP
26. On-Chip Memory II (RAM or ROM) Intel FPGA IP
27. Optrex 16207 LCD Controller Core
28. PIO Core
29. PLL Cores
30. DMA Controller Core
31. Modular Scatter-Gather DMA Core
32. Scatter-Gather DMA Controller Core
33. SDRAM Controller Core
34. Tri-State SDRAM Core
35. Video Sync Generator and Pixel Converter Cores
36. Intel FPGA Interrupt Latency Counter Core
37. Performance Counter Unit Core
38. Vectored Interrupt Controller Core
39. Avalon® -ST Data Pattern Generator and Checker Cores
40. Avalon® -ST Test Pattern Generator and Checker Cores
41. System ID Peripheral Core
42. Avalon® Packets to Transactions Converter Core
43. Avalon® -ST Multiplexer and Demultiplexer Cores
44. Avalon® -ST Bytes to Packets and Packets to Bytes Converter Cores
45. Avalon® -ST Delay Core
46. Avalon® -ST Round Robin Scheduler Core
47. Avalon® -ST Splitter Core
48. Avalon® -MM DDR Memory Half Rate Bridge Core
49. Intel FPGA GMII to RGMII Converter Core
50. Intel FPGA MII to RMII Converter Core
51. HPS GMII to TSE 1000BASE-X/SGMII PCS Bridge Core Intel® FPGA IP
52. Intel FPGA HPS EMAC to Multi-rate PHY GMII Adapter Core
53. Intel FPGA MSI to GIC Generator Core
54. Cache Coherency Translator Intel® FPGA IP
55. Lightweight UART Core
10.2.1. Unsupported Features
10.2.2. Interface
10.2.3. General Architecture
10.2.4. 16550 UART General Programming Flow Chart
10.2.5. Configuration Parameters
10.2.6. DMA Support
10.2.7. FPGA Resource Usage
10.2.8. Timing and Fmax
10.2.9. Avalon® -MM Agent
10.2.10. Over-run/Under-run Conditions
10.2.11. Hardware Auto Flow-Control
10.2.12. Clock and Baud Rate Selection
15.5.2.1. Transfer Command FIFO (TFR_CMD)
15.5.2.2. Receive Data FIFO (RX_DATA)
15.5.2.3. Control Register (CTRL)
15.5.2.4. Interrupt Status Enable Register (ISER)
15.5.2.5. Interrupt Status Register (ISR)
15.5.2.6. Status Register (STATUS)
15.5.2.7. TFR CMD FIFO Level (TFR CMD FIFO LVL)
15.5.2.8. RX Data FIFO Level (RX Data FIFO LVL)
15.5.2.9. SCL Low Count (SCL LOW)
15.5.2.10. SCL High Count (SCL HIGH)
15.5.2.11. SDA Hold Count (SDA HOLD)
24.6.1. altera_avalon_fifo_init()
24.6.2. altera_avalon_fifo_read_status()
24.6.3. altera_avalon_fifo_read_ienable()
24.6.4. altera_avalon_fifo_read_almostfull()
24.6.5. altera_avalon_fifo_read_almostempty()
24.6.6. altera_avalon_fifo_read_event()
24.6.7. altera_avalon_fifo_read_level()
24.6.8. altera_avalon_fifo_clear_event()
24.6.9. altera_avalon_fifo_write_ienable()
24.6.10. altera_avalon_fifo_write_almostfull()
24.6.11. altera_avalon_fifo_write_almostempty()
24.6.12. altera_avalon_write_fifo()
24.6.13. altera_avalon_write_other_info()
24.6.14. altera_avalon_fifo_read_fifo()
24.6.15. altera_avalon_fifo_read_other_info()
25.1. Core Overview
25.2. Component-Level Design for On-Chip Memory
25.3. Platform Designer System-Level Design for On-Chip Memory
25.4. Simulation for On-Chip Memory
25.5. Intel® Quartus® Prime Project-Level Design for On-Chip Memory
25.6. Board-Level Design for On-Chip Memory
25.7. Example Design with On-Chip Memory
25.8. On-Chip Memory (RAM and ROM) Intel FPGA IP Revision History
26.1. Core Overview
26.2. Embedded Memory Architecture and Features
26.3. Component-Level Configurations
26.4. Interface Signals
26.5. Control and Status Registers
26.6. Software Programming Model
26.7. Platform Designer System-Level Design for On-Chip Memory II
26.8. Simulation for On-Chip Memory II
26.9. Intel® Quartus® Prime Project-Level Design for On-Chip Memory II
26.10. Board-Level Design for On-Chip Memory II
26.11. Example Design with On-Chip Memory II
26.12. On-Chip Memory II (RAM and ROM) Intel FPGA IP Revision History
31.1. Core Overview
31.2. Feature Description
31.3. mSGDMA Interfaces and Parameters
31.4. mSGDMA Descriptors
31.5. Register Map of mSGDMA
31.6. Programming Model
31.7. Modular Scatter-Gather DMA Prefetcher Core
31.8. Driver Implementation
31.9. Example Code Using mSGDMA Core
31.10. Modular Scatter-Gather DMA Core Revision History
31.5.1. Status Register
31.5.2. Control Register
31.5.3. Write Fill Level Register
31.5.4. Read Fill Level Register
31.5.5. Response Fill Level Register
31.5.6. Write Sequence Number Register
31.5.7. Read Sequence Number Register
31.5.8. Component Configuration 1 Register
31.5.9. Component Configuration 2 Register
31.5.10. Component Type Register
31.5.11. Component Version Register
31.8.1. alt_msgdma_standard_descriptor_async_transfer
31.8.2. alt_msgdma_extended_descriptor_async_transfer
31.8.3. alt_msgdma_descriptor_async_transfer
31.8.4. alt_msgdma_standard_descriptor_sync_transfer
31.8.5. alt_msgdma_extended_descriptor_sync_transfer
31.8.6. alt_msgdma_descriptor_sync_transfer
31.8.7. alt_msgdma_construct_standard_st_to_mm_descriptor
31.8.8. alt_msgdma_construct_standard_mm_to_st_descriptor
31.8.9. alt_msgdma_construct_standard_mm_to_mm_descriptor
31.8.10. alt_msgdma_construct_standard_descriptor
31.8.11. alt_msgdma_construct_extended_st_to_mm_descriptor
31.8.12. alt_msgdma_construct_extended_mm_to_st_descriptor
31.8.13. alt_msgdma_construct_extended_mm_to_mm_descriptor
31.8.14. alt_msgdma_construct_extended_descriptor
31.8.15. alt_msgdma_register_callback
31.8.16. alt_msgdma_open
31.8.17. alt_msgdma_write_standard_descriptor
31.8.18. alt_msgdma_write_extended_descriptor
31.8.19. alt_msgdma_init
31.8.20. alt_msgdma_irq
32.7.1. Data Structure
32.7.2. SG-DMA API
32.7.3. alt_avalon_sgdma_do_async_transfer()
32.7.4. alt_avalon_sgdma_do_sync_transfer()
32.7.5. alt_avalon_sgdma_construct_mem_to_mem_desc()
32.7.6. alt_avalon_sgdma_construct_stream_to_mem_desc()
32.7.7. alt_avalon_sgdma_construct_mem_to_stream_desc()
32.7.8. alt_avalon_sgdma_enable_desc_poll()
32.7.9. alt_avalon_sgdma_disable_desc_poll()
32.7.10. alt_avalon_sgdma_check_descriptor_status()
32.7.11. alt_avalon_sgdma_register_callback()
32.7.12. alt_avalon_sgdma_start()
32.7.13. alt_avalon_sgdma_stop()
32.7.14. alt_avalon_sgdma_open()
38.5.6.1. altera_vic_driver.enable_preemption
38.5.6.2. altera_vic_driver.enable_preemption_into_new_register_set
38.5.6.3. altera_vic_driver.enable_preemption_rs_<n>
38.5.6.4. altera_vic_driver.linker_section
38.5.6.5. altera_vic_driver.<name>.vec_size
38.5.6.6. altera_vic_driver.<name>.irq<n>_rrs
38.5.6.7. altera_vic_driver.<name>.irq<n>_ril
38.5.6.8. altera_vic_driver.<name>.irq<n>_rnmi
38.5.6.9. Default Settings for RRS and RIL
38.5.6.10. VIC BSP Design Rules for Intel FPGA HAL Implementation
38.5.6.11. RTOS Considerations
40.1. Core Overview
40.2. Resource Utilization and Performance
40.3. Test Pattern Generator
40.4. Test Pattern Checker
40.5. Hardware Simulation Considerations
40.6. Software Programming Model
40.7. Test Pattern Generator API
40.8. Test Pattern Checker API
40.9. Avalon® -ST Test Pattern Generator and Checker Cores Revision History
40.7.1. data_source_reset()
40.7.2. data_source_init()
40.7.3. data_source_get_id()
40.7.4. data_source_get_supports_packets()
40.7.5. data_source_get_num_channels()
40.7.6. data_source_get_symbols_per_cycle()
40.7.7. data_source_set_enable()
40.7.8. data_source_get_enable()
40.7.9. data_source_set_throttle()
40.7.10. data_source_get_throttle()
40.7.11. data_source_is_busy()
40.7.12. data_source_fill_level()
40.7.13. data_source_send_data()
40.8.1. data_sink_reset()
40.8.2. data_sink_init()
40.8.3. data_sink_get_id()
40.8.4. data_sink_get_supports_packets()
40.8.5. data_sink_get_num_channels()
40.8.6. data_sink_get_symbols_per_cycle()
40.8.7. data_sink_set enable()
40.8.8. data_sink_get_enable()
40.8.9. data_sink_set_throttle()
40.8.10. data_sink_get_throttle()
40.8.11. data_sink_get_packet_count()
40.8.12. data_sink_get_symbol_count()
40.8.13. data_sink_get_error_count()
40.8.14. data_sink_get_exception()
40.8.15. data_sink_exception_is_exception()
40.8.16. data_sink_exception_has_data_error()
40.8.17. data_sink_exception_has_missing_sop()
40.8.18. data_sink_exception_has_missing_eop()
40.8.19. data_sink_exception_signalled_error()
40.8.20. data_sink_exception_channel()
Visible to Intel only — GUID: qcm1488476882488
Ixiasoft
25.2.2. Size
This options defines the size and width of the memory.
- Enable different width for Dual-port Access—Different width for dual-port access status.
Note: A different width for dual-port access is not supported for Intel® Stratix® 10 devices.
- Agent S1 Data width—This setting determines the data width of the memory. The available choices are 8, 16, 32, 64, 128, 256, 512, or 1024 bits. Assign Data width to match the width of the host that accesses this memory the most frequently or has the most critical throughput requirements. For example, if you are connecting the on-chip memory to the data host of a Nios® II or Nios® V processor, you should set the data width of the on-chip memory to 32 bits, the same as the data-width of the Nios® II or Nios® V processor data host. Otherwise, the access latency could be longer than one cycle because the Avalon® interconnect fabric performs width translation.
- Total memory size—This setting determines the total size of the on-chip memory block. The total memory size must be less than the available memory in the target FPGA.
The IP parameter editor accepts characters k and m to specify the memory size in kilobytes and megabytes respectively. For example: if you enter 1k, it will automatically resolve to its equivalent bytes which is 1024 bytes in this case.
- This option is only available for devices that include M4K memory blocks. If selected, the Intel® Quartus® Prime software divides the memory by depth rather than width, so that fewer memory blocks are used. This change may decrease fmax.