Numerical determination of the distance between the GPUs where effective cooling is made in computers used in coin mining


Onder A. E. , ULUSARSLAN D.

NUMERICAL HEAT TRANSFER PART A-APPLICATIONS, 2022 (Peer-Reviewed Journal) identifier identifier

  • Publication Type: Article / Article
  • Publication Date: 2022
  • Doi Number: 10.1080/10407782.2022.2063574
  • Journal Name: NUMERICAL HEAT TRANSFER PART A-APPLICATIONS
  • Journal Indexes: Science Citation Index Expanded, Scopus, Academic Search Premier, Aerospace Database, Communication Abstracts, Compendex, INSPEC, Metadex, Civil Engineering Abstracts
  • Keywords: Axial fans, convection, cooling, GPU, heat transfer, numerical analysis, PLATES, ARRAY

Abstract

Computers need more and more powerful processors and GPUs for complex processes such as crypto operations, which makes the cooling more critical. The air flow rate created by axial fans used in cooling may widely differ depending on the resistance, which is nearly the same at their force ends, but the other components, such as another GPU, installed near their lift ends differs considerably. In this study, experimental and numeric investigations were performed on operating points of fans due to the resistance caused by other GPUs and its effect on cooling process. The longer the distance between the GPUs, the less the GPUs warmed in the direction of the air flow, the less cooling air flow speed, and the less convection heat transfer coefficients. The optimal distance for best cooling performance were found out. It was seen that the results of the experimental and numeric analyses performed were in good agreement.