I am also from India, and I checked all the dealers and websites, nobody was offering me a 2080ti for less than 1.25 lacs. You got lucky there, mate. 🙂
@Den-Fi Does the UPS shutoff power to the CPU if it gets overloaded while its connected to current from wall? I really don't mind about backup purposes, I am worried about voltage fluctuations at my place and them not damaging my CPU.
F4-3200C14Q-32GFX is a Samsung B-die. Flare X kits usually are well compatible with x399 platform so you should be good to go. Let us know once you test them!
hey @timbergus are you running a XMP profile with 3200Mhz or more? Stock setting will run at 2133Mhz. In My X399 Gaming Carbon AC for a 3200Mhz Ram it gives XMP1 as 2933 and XMP2 as 3200. Though for x399, you should manually tune the timings to achieve best settings. I was able to run 3133Mhz with 16-18-18-38 manually with my Hynix ripjaw 3200Mhz chip.
And from what I have read, max speed people could run with 8 Rams in threadripper is 2933 Mhz (Correct me anyone if I am wrong). With 2 or 4 rams they could achieve upto 3600Mhz which is what I am looking at.
Also, your RAM series is 18D, which is rated as a dual channel RAM and you are mixing 4 kits, so it is 50/50 chance if it can run over 2933Mhz. (Which was my concern as well with F4-3600C15D-16GTZ as it is a dual channel RAM and not sure how high it can go with Quad channel). You can read more here: Don't mix Kits
Coming to latency, lesser response time is better performing in multi-core applications and in benchmarks as well, but in gaming its random as per what I've seen (depends on game). Though low latency is always preferable.
With these calculations the best timings for ryzen/threadripper is: 3600CL15 > 3200CL14 > 3600CL16
Response time you can calculate as below:
CL / frequency x 1000 = ns (response time)
some examples
cl14 / 3200 x 1000 = 4.375ns
cl15 / 3466 x 1000 = 4.328ns
cl18 / 3600 x 1000 = 5ns
cl19 / 4000 x 1000 = 4.75ns