Heat and the CPU die size

This content was originally featured on Amdmb.com and has been converted to PC Perspective’s website. Some color changes and flaws may appear.

Most users are familiar with the major trends occurring in the computer industry as related to CPU speed, heat output and die size. Processors are constantly being designed and refined to run at ever-higher clock frequencies. CPUs do no physical work so nearly all of the electrical energy that flows into a processor comes back out in the form of thermal energy. The amount of heat generated is determined by a number of factors including, design, fabrication technique, operating voltage, clock frequency and application load. Another trend we continually see is the shrinking size of the semiconductor chip or die. Fabricators keep moving to smaller and smaller die sizes, driven by better performance and economics. One big advantage of smaller, denser circuits is that they can operate faster and at lower voltages. And lower voltage generally means less heat production.

The big thing many people do not realize though is that as the CPU die size shrinks the available surface area to dissipate heat into the heatsink base also gets smaller! For example:

AMD Processor

Manufact. Process

Die size

Heat output max.

Thunderbird 1.3

0.18 micron

120 mm^2

68.3 watts

Thoroughbred 2400+

0.13 micron

  84 mm^2

68.3 watts

As you can see when we compare an old Thunderbird CPU to a newer Thoroughbred chip, the same amount of thermal energy now has to flow thru a significantly smaller area. This means that the actual area of contact between the CPU and heatsink base plate has been reduced by 30%.

The point to all this, is to emphasize the importance of the heat transfer capabilities of the heatsink base. Both the Cooler Master XDream SE and Aero 7 incorporate an all copper, skived fin heatsink. Pure copper has the highest thermal conductivity available next to silver and diamond.

Skiving – the new buzz in heatsink manufacturing

More and more high-performance HSF manufacturers are using skiving technology to produce heatsinks. “So what’s all the fuss about skiving?” you ask. Skiving is a process used to manufacture a heatsink from a solid block of material. Most heatsinks incorporate numerous metal fins or pins, which are either soldered or pressed into the base plate. The joints formed by mechanical bonding create tiny barriers to the flow of heat. Many solid aluminum heatsinks are already manufactured using the extrusion process. Copper is heavier and more expensive than aluminum but conducts heat better. Unfortunately copper cannot be extruded into finished heatsink shapes the way aluminum can. Copper is also relatively difficult to machine so for reasons of manufacturability and cost, most copper heatsinks have been built up from many smaller pieces.

Cooler Master: The XDream SE and Aero 7 - Cases and Cooling 15
Cooler Master skived fin construction

During the skiving process, material is removed by sharp carbide cutters in a linear fashion; similar to planing. Typically the cutters are stationary while the work piece moves back and forth, shaving off a new layer of material with each pass. Skiving produces an excellent surface finish and does not deform or work harden the material from heat, as do other machining processes like milling. Because of this, much thinner fins can be produced and with closer spacing. The skiving process offers several major advantages:

  • Heatsink can be carved from a single, homogeneous block of metal
  • Fins can be made thinner
  • Fins can be spaced closer together
  • Smooth surface finish with no work hardening
Put these all together and you potentially have a heatsink with lower thermal resistance and higher surface area than a traditional built-up copper heatsink can provide.
« PreviousNext »