[10/10/10/10/10/20] <1.5,1.9> General-purpose processes are optimized for general-purpose computing. That is, they are optimized for behavior that is gener- ally found across a large number of applications. However, once the domain is restricted somewhat, the behavior that is found across a large number of the target applications may be different from general-purpose applications. One such appli- cation is deep learning or neural networks. Deep learning can be applied to many different applications, but the fundamental building block of inferenceusing the learned information to make decisionsis the same across them all. Inference operations are largely parallel, so they are currently performed on graphics proces- sing units, which are specialized more toward this type of computation, and not to inference in particular. In a quest for more performance per watt, Google has cre- ated a custom chip using tensor processing units to accelerate inference operations in deep learning.1 This approach can be used for speech recognition and image recognition, for example. This problem explores the trade-offs between this pro- cess, a general-purpose processor (Haswell E5-2699 v3) and a GPU (NVIDIA K80), in terms of performance and cooling. If heat is not removed from the com- puter efficiently, the fans will blow hot air back onto the computer, not cold air. Note: The differences are more than processoron-chip memory and DRAM also come into play. Therefore statistics are at a system level, not a chip level. If Googles data center spends 70% of its time on workload A and 30% of its time on workload B when running GPUs, what is the speedup of the TPU system over the GPU system? b. [10] < 1.9> If Googles data center spends 70% of its time on workload A and 30% of its time on workload B when running GPUs, what percentage of Max IPS does it achieve for each of the three systems? c. [15] < 1.5, 1.9> Building on (b), assuming that the power scales linearly from idle to busy power as IPS grows from 0% to 100%, what is the performance per watt of the TPU system over the GPU system? d. [10] < 1.9> If another data center spends 40% of its time on workload A, 10% of its time on workload B, and 50% of its time on workload C, what are the speedups of the GPU and TPU systems over the general-purpose system? e. [10] < 1.5> A cooling door for a rack costs$4000 and dissipates 14 kW (into the room; additional cost is required to get it out of the room). How many Haswell-, NVIDIA-, or Tensor-based servers can you cool with one cooling door, assuming TDP inFigures 1.27and1.28? f. [20] < 1.5> Typical server farms can dissipate a maximum of 200 W per square foot. Given that a server rack requires 11 square feet (including front and back clearance), how many servers from part (e) can be placed on a single rack, and how many cooling doors are required? System Chip Throughput % Max IPS A B C ABC General-purpose Haswell E5-2699 v3 5482 13,194 12,000 42% 100% 90% Graphics processor NVIDIA K80 13,461 36,465 15,000 37% 100% 40% Custom ASIC TPU 225,000 280,000 2000 80% 100% 1% Figure 1.28 Performance characteristics for general-purpose processor, graphical processing unit-based or custom ASIC-based system on two neural-net workloads (cite ISCA paper). Workloads A and B are from published results. Workload C is a fictional, more general-purpose application. System Chip TDP Idle power Busy power General-purpose Haswell E5-2699 v3 504 W 159 W 455 W Graphics processor NVIDIA K80 1838 W 357 W 991 W Custom ASIC TPU 861 W 290 W 384 W Figure 1.27 Hardware characteristics for general-purpose processor, graphical processing unit-based or custom ASIC-based system, including measured power (cite ISCA paper).