Workgroups of waves are dispatched in one of two modes: CU or WGP. This mode controls whether the waves of a workgroup are distributed across just two SIMD32’s (CU mode), or across all 4 SIMD32’s (WGP mode) within a WGP. In CU mode, waves are allocated to two SIMD32’s which share a texture memory unit, and are allocated LDS space which is all local (on the same side) as the SIMDs. This mode can provide higher LDS memory bandwidth than WGP mode. In WGP mode, the waves are distributed over all 4 SIMD32’s and LDS space maybe allocated anywhere within the LDS memory. Waves may access data on the "near" or "far" side of LDS equally, but performance may be lower in some cases. This mode provides more ALU and texture memory bandwidth to a single workgroup (of at least 4 waves).
LDS Access
There are 3 forms of Local Data Share access: • Direct Read – reads a single dword from LDS and broadcasts the data as input to a vector ALU op. • Indexed Read/write and Atomic ops – read/write address comes from a VGPR and data to/from VGPR. ◦ LDS-ops require up to 3 inputs: 2data+1addr and immediate return VGPR. • Parameter Interpolation – similar to direct read but with specific addressing. ◦ Reads up to 2 parameters (P0, P1-P0) or (P2-P0) from one attribute to be supplied to a muladd. ◦ Also supplies individual parameter read for general interpolation (or select I,J=0.0) The following subsections describe these methods.
LDS Direct Read
Direct reads are only available in LDS, not in GDS. LDS Direct reads occur in vector ALU (VALU) instructions and allow the LDS to supply a single DWORD value which is broadcast to all threads in the wavefront and is used as the SRC0 input to the ALU operations. A VALU instruction indicates that input is to be supplied by LDS by using the LDS_DIRECT for the SRC0 field. The LDS address and data-type of the data to be read from LDS comes from the M0 register
Our superior products
ABB -- AC 800M controller, Bailey, PM866 controller, IGCT silicon controlled 5SHY 3BHB01 3BHEO0 3HNA00 DSOC series
BENTLY --- 3500 system/proximitor, front and rear cards, sensors, power modules, probes, cables
Emerson -- modbus card, power panel, controller, power supply, base, power module, switch
EPRO --- Data acquisition module, probe, speed sensor, vibration sensor, shaft vibration transmitter, proximitor
FOXBORO - thermal resistance input/output module, power module, communication module, cable, controller, switch
GE --- module, air switch, I/O module, display, CPU module, power module, converter, CPU board, Ethernet module, integrated protection device, power module, gas turbine card
HIMA --- DI module, processor module, AI card, pulse encoder
Honeywell --- Secure digital output card, program module, analog input card, CPU module, FIM card
MOOG - servo valve, controller, module, power module
NI --- Information acquisition card, PXI module, card, chassis multi-channel control card
WESTINGHOUSE --- RTD thermal resistance input module, AI/AO/DI/DO module, power module, control module, base module
Woodward - Regulator, module, controller, governor
YOKOGAWA - Servo module, control cabinet node unit