Building a workstation: i7 vs Xeon CPUs

Before building a computer you’ve got to ask yourself one important question: What am I going to use it for?


Now, if you’re reading this then you know what I do requires a lot of CPU horsepower so some dinky little processor isn’t going to cut it. That’s why we’ve got to look to the high end I7 and Xeon processors.


Think of Intel i7 CPUs as high end consumer chips that you can buy in a shop. They’re flexible (in what they can be used for) and contain a lot of features. The Xeons however are enterprise level chips that are good at doing lots of things at the same time… but only if you have software that can take advantage of it.

As it stands a lot of software simply can’t scale up to accommodate a 48 core monstrosity (or really anything above 8 cores if that) because programming multithreaded software is a difficult task.

MORE POWER = MORE POWER (I mean electricity) + HEAT

In addition each core in those monstrous multicore CPUs runs at increasingly lower speeds the more cores are added. So if your 4 core I7 runs at 4GHz then your 24 core Xeon will be operating in the region of 2.2 GHz per core. That’s because if they were much faster it would generate so much heat that the entire setup would burn hotter than the sun and ignite our atmosphere.

That or it would fry and then die before you could use it.

But if you’ve got the software that can make use of those slower cores then that Xeon will significantly outperform the i7. If not then most of those cores will be sitting there twiddling their proverbial digital thumbs.

So much for £5,000 you spent on those top end dual Xeons.

“But I could just overclock the Xeons!” Xeons are often locked at a frequency because that was what they were designed to handle. In some instances you can overclock them slightly but they are notoriously unstable at higher clock frequencies.

Only do it if you have money to burn, because quite honestly you’re going to fry your CPU if you try.


Sorenson Squeeze is a video encoding application that can operate multiple instances simultaneously. That means you could simultaneously transcode multiple clips of raw footage into prores proxies (assuming that’s what your workflow demanded) using a lot of the cores while having spares left to process whatever other demands you place on the CPU.

So if you can make use of that power then you’ll get your money’s worth. But what you get in massively multicore performance you’ll lose in flexibility and features.

The alternative option is the extreme I7s that come in 6 and 8 core varieties. With hyperthreading enabled (2 threads per core) you can get Xeon-like multicore performance in a CPU that connects to consumer motherboards. That means that you get access to a lot of nice features that otherwise are omitted in no-frills xeon systems.


GPU acceleration: Processing duties can be offloaded to a fast graphics card. A lot of software makes use of this to speed up transcoding and rendering. Xeon and 6/8 core I7s have a LOT of PCI-E lanes, which means you could run multiple graphics cards simultaneously.

Quicksync: Encoding of H264 footage (a common codec used when exporting deliverables) can be offloaded to the graphics core embedded within the CPU. This is a feature not available on the super high-end extreme I7s described above.

ECC memory: A feature common in Xeon CPUs. This stands for error correcting code memory. Basically, it corrects data corruption.


I am a big believer that picking the right tool for the right job is vital in ensuring that the job gets done well.

And better utilisation of processing power makes on-set processing a possibility. This means you can save money by employing me to do in minutes what will take a day at a minimum for large established post-production companies.

So if you want a tailored approach to on-set post then feel free to get in touch.