14 DAY TRIAL // During the ISSCC conference, which takes place right now in San Francisco, California, IBM has detailed its 5.2GHz z196 flagship CPU, which features an 18% higher clock than the previous 4.4GHz z10 processor, while also maintaining a similar thermal envelope.
IBM managed to achieve this feat by moving from 65nm to a 45nm SOI fabrication process which allowed them to go for faster clock speeds while also maintaining the chip's power consumption in check, reports EETimes.
The new processor also makes extensive use of embedded DRAM for memory and features an out-of-order architecture, which allows IT to perform up to 40% better in some applications compared to its predecessor.
"We think there is still room for future improvements, but frequency increases won't go on forever," said Jim Warnock, a senior IBM engineer.
"We are the last of the high-end processors still pushing higher frequency," continued Warnock.
The z196 CPU is used in multichip modules which are installed in IBM's z-Series servers, such a module including six CPUs and two L4 cache chips while its power consumption is estimated at a whopping 1,800 Watts.
After subtracting the power used by the L4 cache, it results that each processor consumes as much as 260 Watt, but the energy efficiency of the modules could be improved if better e MCM (multi-chip module) materials would be available.
The z196 chip implements a CISC z/Architecture and is made up of 1.4 billion transistors, its total die size reaching 512.3mm2.
Each of the four cores comes with 64 KB of L1 instruction cache, a 128 KB L1 data cache and a 1.5 MB L2 cache, while the 24MB L3 cache is shared between all the cores and controlled by two on-chip L3 cache ICs.
The cores feature six RISC-like execution units, including two integer units, two load-store units, one binary floating point unit and one decimal floating point unit, and the CPU also includes a DDR3 RAM memory controller.