Chainin Parameter
The following are exception flags supported in flushed and bfloat16 formats:
•
fp16_mult_top_invalid
•
fp16_mult_top_inexact
•
fp16_mult_top_overflow
•
fp16_mult_top_underflow
•
fp16_mult_bot_invalid
•
fp16_mult_bot_inexact
•
fp16_mult_bot_overflow
•
fp16_mult_bot_underflow
•
fp16_adder_invalid
•
fp16_adder_inexact
•
fp16_adder_overflow
•
fp16_adder_underflow
•
fp32_adder_invalid
•
fp32_adder_inexact
•
fp32_adder_overflow
•
fp32_adder_underflow
The following are exception flags supported in extended format:
•
fp16_mult_top_invalid
•
fp16_mult_top_inexact
•
fp16_mult_top_infinite
•
fp16_mult_top_zero
•
fp16_mult_bot_invalid
•
fp16_mult_bot_inexact
•
fp16_mult_bot_infinite
•
fp16_mult_bot_zero
•
fp16_adder_invalid
•
fp16_adder_inexact
•
fp16_adder_infinite
•
fp16_adder_zero
•
fp32_adder_invalid
®
™
Intel
Agilex
Variable Precision DSP Blocks User Guide
52
3. Intel Agilex Variable Precision DSP Blocks Operational Modes
Vector Two with Floating-Point
Addition
fp32_chainout = (
fp16_mult_top_a
*
) +
fp16_mult_top_b
(
*
fp16_mult_bot_a
)
fp16_mult_bot_b
UG-20213 | 2019.04.02
Vector Two with Floating-Point
Subtraction
fp32_chainout = (
fp16_mult_top_a
*
) -
fp16_mult_top_b
(
*
fp16_mult_bot_a
)
fp16_mult_bot_b
Send Feedback