Exploring the diversity of multimedia systems
read more
Citations
Processor evaluation cube : A classification and survey of processor evaluation techniques
References
Computers and Intractability: A Guide to the Theory of NP-Completeness
Computer Architecture: A Quantitative Approach
The Art of Computer Systems Performance Analysis.
The art of computer systems performance analysis
Related Papers (5)
Frequently Asked Questions (17)
Q2. What is the objective function of the optimizer?
The objective function of the optimizer is minimization of selected machine configurations, thereby maximizing the number of benchmarks that can be run on a processor as though it is optimized for each individual benchmark.
Q3. What is the effect of memory operations on other measurements?
Note that in order to reduce the effect of memory operations on other measurements, the target machine has 32 KB instruction cache and 32 KB data cache, resulting in high cache hit rates.
Q4. What are the run-time characteristics of the machine?
The run-time characteristics include the available ILP, demand on various hardware components such as cache memory units, register files, and the number of function units.
Q5. What is the main idea behind the idea of programmable processors?
The arrival of production quality ILP compilers and commercial DSPs with VLIW architecture stimulated the idea of programmable processors that are aggressively tuned to specific applications.
Q6. how many machine configurations can be run on a processor?
The objective function of the optimization problem is minimization of the number of selected machine configurations, thereby, on average, maximizing the number of benchmarks that can be run on a processor as though it is optimized for each individual benchmark.
Q7. Why did Conte and MangioneSmith use ad hoc methods?
Early work in the area of processor architecture synthesis tended to employ ad hoc methods on small code kernels, in large part due to the lack of good retargetable compiler technology.
Q8. What is the role of cache memory units in machine performance?
The authors incorporate the role of cache memory units in machine performance into the machine model, which is essential for producing meaningful results.
Q9. How many mm2 of area is needed for the given compiler technology?
for the given compiler technology and benchmarks, there is no need to have more than 100 mm2 of area since the speed-up increase achieved by machines greater than 100 mm2 are minimal.
Q10. what is the framework used in the design of a chip?
The authors have found that the framework introduced in this paper can be very valuable in making early design decisions such as area and architectural configuration tradeoff, cache and issue width tradeoff under area constraint, and the number of branch units and issue width.
Q11. How can the authors estimate the issue unit area of a superscalar machine?
For a superscalar machine, the issue unit area cannot be estimated using a simple linear model since it requires more complex logic for runtime code scheduling.
Q12. How many applications are used in this study?
The collection is composed of 21 applications culled from available image processing, communications, cryptography, and DSP applications.
Q13. How many byte-sized instruction caches are used in the benchmark?
For each executable of a benchmark, the authors simulate 25 combina-tions of instruction cache and data cache ranging from (512 bytes, 512 bytes) to (8 KB, 8 KB).
Q14. What is the reason why the ILP is not fully exploited?
One of the underlying reasons that causes the phenomenon is that the ILP found by the compiler and hardware scheduler is fully exploited by having a certain amount of hardware, thereby performance increase possibility is exhausted.
Q15. How many mm2 of area are available for the given compiler technology?
for the given compiler technology and benchmarks, there is little need to have more than 100 mm2 of area sincethe speed-up increase achieved by machines greater than 100 mm2 are minimal.
Q16. What are the parameters of the Lsim simulator?
Memory latency, misprediction penalty and ALU latency are specified as Lsim parameters (Fig. 2) in the system model being simulated.
Q17. What are the characteristics of the benchmarks?
Note that the combination of the instructions per cycle (IPC), bus utilization, branch issue, and ALU issue exhibit distinctive characteristics for each benchmark.