ajc9988
Senior member
- Apr 1, 2015
- 278
- 171
- 116
No, testing at the same power DOES NOT give IPC. IPC is a DEFINED term. Do you not understand that defined terms and how they are tested are what they are.Lack of understanding you said...?..
What is the purpose of comparing at same frequencies other than to check the arch logical efficency without looking at the electrical efficency..?.
Because there s no use of a CPU hat would have 10% higher IPC at the expense of 20% more power.
You think that say 10% more throughput/Hz is free of power charge..?
So testing at the same power will give an indication of IPC and intrinsical efficency of the uarch, ultimately the only thing that mater is the perf/watt, exactly what i m stating, and in this respect one has to admit that 10% better Cinebench score at same 15W power than a 12nm based product is nothing to bragg about, looking at Intel vs Intel 10nm ICL has 20% better perf (and perf/watt) at isopower than the most refined 14nm SKL parts.
You say such a measurement is useless. That's fine and is your opinion. But what you cannot say is that measuring at a fixed power draw measures IPC! IT DOES NOT! Instead, it gives the performance at a specific power draw, which is performance per watt. THAT IS A DIFFERENT METRIC. Sure, it is more apt at determining performance within a specific power draw scenario, which can make it more useful, and is reliant on IPC as a contributing factor that is baked in, but it, in and of itself, does NOT give the IPC number itself.
So you are wrong that it gives IPC, because it doesn't. It is performance per watt, which IPCxFrequency gives the performance, or rather it can tell you how many instructions are performed.
You are NOT talking IPC, you are talking overall performance or performance at a specific power draw. That is not the same as IPC.
So you are mixing measurements, trying to get people off of examining IPC, then pushing a different metric. If you only pushed the different metric as more accurate, that is fine. But so grossly misrepresenting what IPC is IS what I am taking issue with.
From WIkipedia (https://en.wikipedia.org/wiki/Instructions_per_cycle):
Calculation of IPC
The calculation of IPC is done through running a set piece of code, calculating the number of machine-level instructions required to complete it, then using high-performance timers to calculate the number of clock cycles required to complete it on the actual hardware. The final result comes from dividing the number of instructions by the number of CPU clock cycles.
The number of instructions per second and floating point operations per second for a processor can be derived by multiplying the number of instructions per cycle with the clock rate (cycles per second given in Hertz) of the processor in question. The number of instructions per second is an approximate indicator of the likely performance of the processor.
The number of instructions executed per clock is not a constant for a given processor; it depends on how the particular software being run interacts with the processor, and indeed the entire machine, particularly the memory hierarchy. However, certain processor features tend to lead to designs that have higher-than-average IPC values; the presence of multiple arithmetic logic units (an ALU is a processor subsystem that can perform elementary arithmetic and logical operations), and short pipelines. When comparing different instruction sets, a simpler instruction set may lead to a higher IPC figure than an implementation of a more complex instruction set using the same chip technology; however, the more complex instruction set may be able to achieve more useful work with fewer instructions.
Factors governing IPC
A given level of instructions per second can be achieved with a high IPC and a low clock speed (like the AMD Athlon and early Intel's Core Series), or from a low IPC and high clock speed (like the Intel Pentium 4 and to a lesser extent the AMD Bulldozer). Both are valid processor designs, and the choice between the two is often dictated by history, engineering constraints, or marketing pressures.[original research?] However, a high IPC with a high frequency will always give the best performance.
Last edited: