I have had a desktop PC for 1 year, with an i7 4790 cpu, (not the K model) with the Intel stock cooler. I use Core temp to monitor CPU temperatures. I soon found out that anything intensive, took the temperature to above 80 C sometimes flirting with 90. I gather from various HARDWARE forums that this is quite normal with the stock cooler. I was thinking of getting a BETTER cooler but before that I have just tried turning off Turbo Boost in the BIOS. Imagine my surprise when a math benchmark I had written (Freebasic, repeated iterations of y = square root of x) is now running the CPU up to only 60-65 (steady for 30 minute runs) and the performance is only around 3% SLOWER. I am quite pleased with this, not least because the Smart Fan is now keeping the machine super quiet.
20 degrees cooler...with only appro 3% hit in performance...i'd say that's a Home Run...The effect seems less with video encoding in Handbrake, but still a 10 degree difference.10 in celcius is a big difference...As I understood it, Turbo Boost kicks in when the CPU is already at 100% to give a boost in clock speed but it seems like you pay a high price in heat generated for a fairly small speed boost. Maybe with a better cooler I might see not so much difference?

Turbo Boost will boost your CPU from its stock speed of 3.6GHz to a max of 3.8GHz when all 4 cores are loaded, 3.9GHz when 3 cores are loaded, and 4GHz when one or two cores are loaded.  So, depending how threaded your benchmark is and assuming Turbo Boost is kicking in fully, you'll see between 5-10% performance boost.  How "aggressively" the CPU boosts depends not only on the CPU load but also power limits and temperatures, so this can vary depending on your cooler but also on your motherboard.  Out of interest does it maintain a steady frequency or does it jump up and down when under load, i.e. turbo disengaging?  This would mean it's hitting a temperature or power limit.

It's not entirely unusual to see quite a difference in temperatures when the CPU's hitting max turbo boost frequency especially with a stock cooler.  What might also be interesting would be if you had a power monitor, you could see what the actual difference in power consumption is.  I suspect that the stock cooler is at its limits of heat dissipation meaning a small amount of extra heat being generated leads to a disproportionate rise in temperature - a power meter would give you an idea of how much extra power is being used and therefore how much more heat is being generated.  You may well find the difference is relatively small and that it might pay off for you to get a slightly better cooler if the frequency fluctuates under load as above.

Hope this helps Laser and IR temp monitors ahve finally gotten reasonable in price.

Handy to have in the shop. Quote from: Calum on July 21, 2016, 01:53:55 PM

Out of interest does it maintain a steady frequency or does it jump up and down when under load, i.e. turbo disengaging?  This would mean it's hitting a temperature or power limit.