2019
DOI: 10.1109/mm.2019.2897782
|View full text |Cite
|
Sign up to set email alerts
|

BROOM: An Open-Source Out-of-Order Processor With Resilient Low-Voltage Operation in 28-nm CMOS

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
7
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
3
2
2

Relationship

0
7

Authors

Journals

citations
Cited by 21 publications
(7 citation statements)
references
References 5 publications
0
7
0
Order By: Relevance
“…It is followed by CVA6 (formerly named Ariane) [44], whose design has been verified on various FPGA boards and through several tapeouts. The high number of Google Scholar hits for BOOM [18] denotes its academic importance as an OoO processor. The SHAKTI C-Class processor [27] is maintained by a slightly smaller community than BOOM and CVA6 and counts two tapeouts.…”
Section: Analysis Of Risc-v Implementationsmentioning
confidence: 99%
“…It is followed by CVA6 (formerly named Ariane) [44], whose design has been verified on various FPGA boards and through several tapeouts. The high number of Google Scholar hits for BOOM [18] denotes its academic importance as an OoO processor. The SHAKTI C-Class processor [27] is maintained by a slightly smaller community than BOOM and CVA6 and counts two tapeouts.…”
Section: Analysis Of Risc-v Implementationsmentioning
confidence: 99%
“…A Single Error Correction Double Error Detection (SEC-DED) code protects the caches of the SiFive U-series IPs (U54, U74) and SoC (FU540), which utilize the 5-stage in-order Rocket processor core. The BROOM tapeout [2] adds resilience methods to the 7-stage out-of-order BOOMv2 processor. Several techniques tolerate hard bit errors in L1 and L2 caches, which allows an aggressive reduction of the core voltage.…”
Section: Related Workmentioning
confidence: 99%
“…The instruction fetch unit is equipped with complex predictors (e.g., GShare and TAGE). A tapeout in TSMC 28 nm achieved 1.0 GHz and a Coremark of 3.77 per MHz [2], which makes BOOM one of the best performing RISC-V implementations. The BOOM utilizes the nonblocking L1D$ version of the Rocket, hence it is afflicted with the same ECC problems as described above.…”
Section: Rocket and Boom Processor Cores Within Chipyardmentioning
confidence: 99%
See 1 more Smart Citation
“…These structures provide good performance, but fall short in power efficiency [2,26], especially in the post-Moore era. Furthermore, these structures also could affect the clock rate and require custom design which lags the tape-out flow of high-performance processors [10].…”
mentioning
confidence: 99%