Xeons vs i7's Gaming Performance
Xeons vs i7's Gaming Performance
I have a colleague who’s undecided about purchasing a workstation from me for a much cheaper price than a brand new desktop. I own a Dell Precision T5810 or an HP Z640, either with an 8-core E5-2640 v3 at 2.6 GHz or a 6-core E5-1650 v3 at 3.5 GHz, and 32 GB of RAM. In Fire Strike Physics the E5-2640 v3 gets a score of 12,449 while the i7-6700K scores 12,155. My concern is whether a solid gaming experience requires high clock speeds or if more cores and threads at lower frequencies are enough for today’s mainstream games. Thanks.
It seems most games are released instead of just one particular title.
It’s a delicate adjustment, but generally quicker cores make a difference. My 6c/6t 8600K ran significantly faster than my 8c/16t R7 2700X in gaming, even with the latter having 166% more threads. The performance drop is more apparent in single core clocks and IPC. With a CPU like a 3700X, 9900K/9900KS, or 3900X, you’ll get better results because of the many cores and their speed. FireStrike works well for testing, but it doesn’t reflect real-world demands. Just as your Xeon might score higher in Cinebench than the 6700K, gaming usually needs 4-6 cores, sometimes up to 8. At stock speeds, those Xeons won’t stand a chance against the 6700K or newer high-clocked CPUs.
Both options are solid workstation choices, though they aren't designed for gaming. While some titles can leverage more than four cores—even eight at times—the typical workload usually centers around a single core, which can reach near full capacity (around 100%) when the average is closer to 60%. The scheduling isn’t ideal. It’s also worth noting that you’re benchmarking against a 6700k processor running at its stock speed (4GHz), which isn’t as demanding as what it might handle under typical gaming conditions. Expect performance similar to a Ryzen 1st generation setup, or worse if overclocked.
For what purpose are these Xeons? They’re essentially similar to Haswell/Haswell-E at comparable speeds, while the 2640 models run slower. The 1650 could work fine, but it shines more when overclocked—R5 1600 can push it to 3.8-4GHz all cores, whereas a single core tops out at 3.8GHz on these chips.
I only look at how stocks performed between the 1700 and 2640v2 because that’s the only option that’s fixed; otherwise, it wouldn’t be accurate. The 1650v3 is available and should surpass the 1600 even with a solid upgrade.
I was considering the E5-1650 v3 as a better choice for gaming on his workstation. Both systems have locked BIOS configurations. (RIP) I’m just checking if this would keep him in the game for a couple of years until he saves more. He already has a 1080 Ti graphics card. Would the 6-core 1650 be sufficient in the meantime, or should we wait to upgrade?
This setup isn't ideal for a workstation unless your motherboard supports it. The chipset doesn’t allow overclocking. If you manage to get a 1650v3 up to 4.2–4.5Ghz, it should provide strong gaming performance. I use an i7 5820K at 4.2Ghz and it performs similarly under those conditions—no noticeable lag. It runs a bit better at 4.5Ghz, but my cooling isn’t sufficient for that speed; it needs around 1.32V to reach 4.5Ghz, while 4.2Ghz works at 1.2V.