F5F Stay Refreshed Hardware Desktop Which frequency does G.Skill offer: 3600MHz 16-16-16-36 or 16-19-19-39?

Which frequency does G.Skill offer: 3600MHz 16-16-16-36 or 16-19-19-39?

Which frequency does G.Skill offer: 3600MHz 16-16-16-36 or 16-19-19-39?

B
bspickler12345
Junior Member
29
12-18-2016, 03:27 AM
#1
Hello everyone, I've been away from the forum for a while but have been checking it out. I'm planning to assemble a new PC since my current laptop is struggling, and due to security concerns, I'll need reliable performance for remote work for at least a month. Gaming quality matters to me, so tighter timing can boost frame rates noticeably. For work, I'll be running medium-sized Unreal games and programming, which means I need at least 32GB of RAM. The two options I'm considering cost €213 and €325— quite a gap. I've looked for direct comparisons but most reviews use 3200MHz kits, which isn't helpful. I also haven't found any data on latency differences during tasks like compiling or building content, which could affect my choice. Anyone have insights or recommendations on performance expectations between these two models?
B
bspickler12345
12-18-2016, 03:27 AM #1

Hello everyone, I've been away from the forum for a while but have been checking it out. I'm planning to assemble a new PC since my current laptop is struggling, and due to security concerns, I'll need reliable performance for remote work for at least a month. Gaming quality matters to me, so tighter timing can boost frame rates noticeably. For work, I'll be running medium-sized Unreal games and programming, which means I need at least 32GB of RAM. The two options I'm considering cost €213 and €325— quite a gap. I've looked for direct comparisons but most reviews use 3200MHz kits, which isn't helpful. I also haven't found any data on latency differences during tasks like compiling or building content, which could affect my choice. Anyone have insights or recommendations on performance expectations between these two models?

M
MrBearIsCool
Junior Member
2
12-18-2016, 04:07 AM
#2
The word "noticeably" may change depending on context.
M
MrBearIsCool
12-18-2016, 04:07 AM #2

The word "noticeably" may change depending on context.

A
Addycassy
Junior Member
37
12-23-2016, 10:15 AM
#3
16-16-16 is a common failure point (almost every b-die set has identical first three timings). I think a few others might work, but I’m not sure which ones. For GSKILL, it consistently results in b-die. 16-19-19 likely belongs to Hynix or another similar brand. Speed and bandwidth will improve with tighter configurations, of course. No one here can confirm how this affects FPS differences unless they’ve actually tested it (or shared someone else’s results). I’m uncertain about the quality of the binning for that particular kit. It’s a dual-rank configuration, which usually leads to worse overclocking compared to single-rank setups. For example, a 4x8GB configuration will perform better than a 2x16GB one. My current setup uses a 3200 C14 with two 16GB modules; it can reach 3600 at 15-15-15-36 at 1.45v. I haven’t tried 16-16-16-36 at that speed, though. It would be cool to see if your model can hit 3600 at 15-15-36 with a lower voltage than 1.45v. Someone reported a 1T dual-rank setup running on an Apex XI, but it was very unstable. Also, I haven’t seen any motherboard that reliably supports dual-rank overclocking above 3466 MHz. Most drop to POSTing around 3333-3400 MHz, so you’d need a 2T configuration then. A 1T chip still offers better performance if it works.
A
Addycassy
12-23-2016, 10:15 AM #3

16-16-16 is a common failure point (almost every b-die set has identical first three timings). I think a few others might work, but I’m not sure which ones. For GSKILL, it consistently results in b-die. 16-19-19 likely belongs to Hynix or another similar brand. Speed and bandwidth will improve with tighter configurations, of course. No one here can confirm how this affects FPS differences unless they’ve actually tested it (or shared someone else’s results). I’m uncertain about the quality of the binning for that particular kit. It’s a dual-rank configuration, which usually leads to worse overclocking compared to single-rank setups. For example, a 4x8GB configuration will perform better than a 2x16GB one. My current setup uses a 3200 C14 with two 16GB modules; it can reach 3600 at 15-15-15-36 at 1.45v. I haven’t tried 16-16-16-36 at that speed, though. It would be cool to see if your model can hit 3600 at 15-15-36 with a lower voltage than 1.45v. Someone reported a 1T dual-rank setup running on an Apex XI, but it was very unstable. Also, I haven’t seen any motherboard that reliably supports dual-rank overclocking above 3466 MHz. Most drop to POSTing around 3333-3400 MHz, so you’d need a 2T configuration then. A 1T chip still offers better performance if it works.

J
Jayden32805
Member
212
12-24-2016, 09:46 AM
#4
As I reflect more, without exact numbers comparing the two delays, I lean toward using approximate values. I’m curious about your comments regarding dual-rank RAM. I’ve read that it can put extra strain on the memory controller, so if I ever need more memory, adding a duplicate might become challenging. How do you determine if the RAM is dual-rank? It’s clear it supports both directions, but figuring out its rank isn’t straightforward. I don’t intend to overclock the RAM myself; I saw on Newegg reviews that G.Skill sometimes recommends raising the voltage slightly to fix timing problems.
J
Jayden32805
12-24-2016, 09:46 AM #4

As I reflect more, without exact numbers comparing the two delays, I lean toward using approximate values. I’m curious about your comments regarding dual-rank RAM. I’ve read that it can put extra strain on the memory controller, so if I ever need more memory, adding a duplicate might become challenging. How do you determine if the RAM is dual-rank? It’s clear it supports both directions, but figuring out its rank isn’t straightforward. I don’t intend to overclock the RAM myself; I saw on Newegg reviews that G.Skill sometimes recommends raising the voltage slightly to fix timing problems.

M
Moritania
Member
56
12-27-2016, 05:26 PM
#5
In fact, activating the XMP profile on either of those kits would produce nearly the same results in most situations.
M
Moritania
12-27-2016, 05:26 PM #5

In fact, activating the XMP profile on either of those kits would produce nearly the same results in most situations.

D
DanyBruinsma
Member
73
12-28-2016, 12:38 AM
#6
I discovered a Ripjaw kit matching those details—its RAM fits both my motherboard and vice versa. It’s nearly half the cost of the B-dies and €35 less than the Trident Neo. From what I’ve learned, the whole “Ryzen optimized” claim seems more about marketing than real performance. Thoughts on this option? Check out the specs here: https://www.gskill.com/specification/165...cification
D
DanyBruinsma
12-28-2016, 12:38 AM #6

I discovered a Ripjaw kit matching those details—its RAM fits both my motherboard and vice versa. It’s nearly half the cost of the B-dies and €35 less than the Trident Neo. From what I’ve learned, the whole “Ryzen optimized” claim seems more about marketing than real performance. Thoughts on this option? Check out the specs here: https://www.gskill.com/specification/165...cification

F
fuhqing
Member
180
12-28-2016, 01:28 PM
#7
I just wanted to check in. I received the Hynix 16-19-19-39 package and set up the XMP 1.35v 3600MHz profile on my Asus X570-P. It’s running at 1T with the 32GB kit, and everything worked perfectly right out of the box—no extra work needed. I’m happy I didn’t invest in the upgrade since it doesn’t seem like it’s worth it anymore.
F
fuhqing
12-28-2016, 01:28 PM #7

I just wanted to check in. I received the Hynix 16-19-19-39 package and set up the XMP 1.35v 3600MHz profile on my Asus X570-P. It’s running at 1T with the 32GB kit, and everything worked perfectly right out of the box—no extra work needed. I’m happy I didn’t invest in the upgrade since it doesn’t seem like it’s worth it anymore.