There are a zillion unique stones on the face of the earth. Each with its own shape, color and composition. But there are many rocks and stones that are of great interest to Palentologists or Archeologists. These are stones on which there are fossils or tools made by prehistoric man. Let us consider that latter. A stone that has been chipped by a human to use used as a spear head or as a crude knife to get meat off the bones of the group’s hunt.

Even in the eyes of an amateur archeologist, these rocks, extermely crude creations, stand out clearly as the “creation” of man. But on the other hand, the human body, an ultra complex system, whose workings man understands only a miniscule fraction, is attributed to chance and “evolution”. Man, must be crazy.

I fail to understand where this arrogance springs from. Once man has learnt a bit a discovered a tad, he seems to become proud and seems to have forgotten the shortcomings of his learning. Given all the time and money in this world and all the scientists put together, man cannot today, create a single mosquitoe. Let alone a mosquitoe, a single celled organism is out of his reach. How can mere chance be so smart to create life and advance it so tactfully? Is it not always logical to believe in a higher creative force? Is man always not born with the idea of a creator?

To believe or not to believe, is ones own responsibility. There can’t be perfect justice on earth. It has to happen elsewhere. For a man who killed one person and for a man who killed a hundred, the punishment is the same? This is ridiculous.

A man might behave well in public but might be a monster privately. This is where a need for the belief in an unseen God is required. An unseen God is better than a visible God, because belief is better than fear. So my friends, believe in the unseen. There is much to gain and nothing to lose.

After the quad core product, now Intel is planning to release an 80 core version in another 5 years time! I really wonder how well the 80 core system will work. Things are definitely taking an interesting turn.

Read more here.

There will be no Gigahertz wars. The Megahertz wars are over. It is becoming more and more difficult to increase CPU speed beyond a point. There are many ways in which CPU performance can be increased without increasing its clock speed. For example by increasing its cache memory size or by improving its Instruction set architecture. For example, RISC CPUs exhibit better performance than their RISC counterparts at the same clock speed. You can read more about the Mhz myth here. Talking about wars, the RISC and CISC wars are also seeing an end since CISC CPUs have long since included many RICS features.
A feature added by Intel to one of its recent processors is HyperThreading. This gives the illusion of two processors where there is actually only one. HyperThreading works by duplicating the datapath of the processor while having a single execution core. There are 2 register sets, but only one execution unit. When one thread experiences a cache miss, the other can run in the meanwhile. Operating systems see a HyperThreaded processor as a regular SMP system.

Intel some time ago announced its Dual core procesors based on its Core architecture. You might have noted that the Core 2 Duo processors are clocked at a slower rate than the Pentium 4s available in the market. Have no doubt, the Core 2 Duo processors with slower clock speeds with beat the “faster” Pentium 4s hands down. This is the way forward for CPUs, rather than continuing to fight the Gigahertz wars. Intel has now also come out with four cores with its Core 2 Quad offering. AMD is not far behind. There are limits to overcome in the basic architecture, however. After the Megahertz war is over, it is not straight forward to start a “number of cores” war. As the number of cores on a processor increase, there will be high bus contention. The system bus for these CPUs is still a common resource for all the cores.

The increase in the number of cores does not increase performance of the system in a linear fashion, there are other issues related to caches, for example. Since RAM is a common resource, before reading from a location, a CPU must make sure that more recent contents for that location are not sitting in the cache of some other CPU. So all these CPUs are constantly chatting with one another about such issues, making linear performance increase impossible. These issues become worse as the number of CPUs increase in an SMP system. There are other architectures like NUMA, that make things less worse, but these are just workarounds, like many things in engineering. Take caches for example.

There are now many interesting probleme to solve! Seymour Cray once noted Anyone can build a fast CPU. The trick is to build a fast system.