• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Help - Hyper X Predator 4800 Mhz

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

TacOp280

New Member
Joined
Dec 24, 2020
Hello, happy holidays all, thank you for taking the time to read. Any help would be very much appreciated.

I'm having a lot of trouble getting this hyper X Predator kit working properly: Spec sheet : https://www.kingston.com/dataSheets/HX448C19PB3K2_16.pdf

Unfortunately I have very little experience overclocking ram beyond activating XMP.
This is the first time I encounter an unstable OC with ram (within XMP limits), also the first time I've owned a kit above 4000 mhz.
Ill try to lay out the problem set as clearly as I can.

Main issue
The 1st XMP profile at 4800mhz CL19-26-26 @1.5V. will not boot.


System
Asus Rog Maximus XII Hero WIFI / BIOS: 0804 (supposedly handles up to 4800mhz)
10700k OC @ 5.0 Ghz
Lian Li Galahad IO 240
2x8gb Hyper X Predator 4800 mhz
JEDEC: DDR4-2400 CL17-17-17 @1.2V
XMP Profile #1: DDR4-4800 CL19-26-26 @1.5V
XMP Profile #2: DDR4-4000 CL19-23-23 @1.35V

Zotac RTX 3090
Corsair RM 850x
Windows 10 home 64

Symptoms/Observations
System was very stable before ram swap, no crashes and low temps. (previous ram was a 3600 cl17 kit at 1.35v)
XMP 1st profile will not boot @ 4800mhz, 1.5v
XMP 1st profile boots with downclock to 4600mhz, but system is unstable and crashes after an hour or so of gaming. Either hard lock or complete shutdown.
XMP 2nd profile 4000mhz @ 1.35v boots and seems stable.
Under 1st XMP profile, VCCIO voltage reads 1.52v & System agent @ 1.616v which seems crazy high to me. CPU core is at 1.35v.
Ram are both in proper A2/B2 slots, reseated them, swapped them, no apparent physical damage or discrepancies.
Chipset drivers are up to date.
Asus recommended the 0804 bios. released in Nov'20 for increased dram compatibility. But there is also a 0904 and a beta 1002. also released in Nov. 20


Desired end states
1) downclock to 4600mhz with the same timings if I can just get it stable.
2) downclock to 4200mhz - 4600mhz and tighten the timings however I have no clue which timings and voltages to try.


Usually id look up other case studies/forums that I can find on the net with a similar kit/system. No luck.
I know most recommend dram calculator to find timings but honestly it feels like I'm just guessing.
The system has gone through quite a beating over the last 72hrs trying to make this work. I think its time I reach out.

Once again thanks for reading. Any help or advice would be appreciated.
 

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
That's some fast and expensive ram... you benchmarking with these or???

Sounds like your IMC or board just isn't capable of reaching those clocks. You can try setting the values manually perhaps with a bit less SA and IO as that isn't daily driver level voltage.

Benefits of running that fast of ram is negligible anyway. Ddr4 4000 is smoking fast as is. :)
 
OP
T

TacOp280

New Member
Joined
Dec 24, 2020
I usually only benchmark for stability tests or to confirm performance gains.
Certainly not the way some of you guys/gals OC/benchmark as a hobby for high scores.

The ram is for a gaming rig, I just wanted to go all out for once & so I picked the highest compatible memory.

I think you're right about the limitations. I just never encountered an xmp profile that fails to boot.
Ill try to downclock to 4400mhz CL19-26-26 @1.5V and reduce IO/SA as you recommended.

Are there better timings I could try?
 

BugFreak

Joined
Apr 29, 2010
Location
Central FL
I wouldn't say it is an XMP profile that can't boot but an XMP the rest of the build can't handle. I'd go with a 4000 or so setting that has the lowest timings you can get and call it a day. You really won't notice the difference and 4000 serves me well in games.
 

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
I'd return the memory for some ddr4 4000 and pocket pocket difference. Ddr4 4000 will work on xmp most of the time.
 

Johan45

Benching Team Leader Super Moderator
Joined
Dec 19, 2012
I usually only benchmark for stability tests or to confirm performance gains.
Certainly not the way some of you guys/gals OC/benchmark as a hobby for high scores.

The ram is for a gaming rig, I just wanted to go all out for once & so I picked the highest compatible memory.

I think you're right about the limitations. I just never encountered an xmp profile that fails to boot.
Ill try to downclock to 4400mhz CL19-26-26 @1.5V and reduce IO/SA as you recommended.

Are there better timings I could try?

We recently tested a DDR4 5000 MHz kit here which might be worth a read https://www.overclockers.com/xpg-spectrix-d50-extreme-ddr4-elite-memory-for-serious-enthusiasts/
Here's an excerpt from the article about trying to get the 5000 MHz XMP to work.
XMP Profile 1 ~ 5000 MHz CL19-28-28
Without touching any of the other settings, simply applying the XMP Profile 1 failed to work. In order to make it work so we could test it, we needed to tweak some settings. The first and most important task was to manually apply the secondary XMP timings. When you apply the XMP profile it only sets the primary timings and VDIMM. However, if you look at the Thaiphoon Burner screenshot above, you will see there are many other timings involved in the XMP profile, and the motherboard does not set automatically.

Here is what we did to make the XMP profile work:

Tested 6 different 10900K CPUs to find the one with the best memory controller for high-frequency memory overclocking.
Manually set all 8 XMP timings, note that some of the timings specified were too high for our motherboard to set, so we went with the maximum value for those.
VDIMM increase from 1.60 V to 1.65 V, which was verified with a digital multimeter
System Agent voltage increase from auto to 1.60 V (1.30 V is considered ’24/7 safe’ in the overclocking community)
IO voltage increase from auto to 1.40 V (1.25 V is considered ’24/7 safe’ in the overclocking community)
Decrease the DRAM termination voltage from 0.825 to 0.750 V
Placed a 2000+ RPM fan directly on the memory to help with cooling
By torturing the CPU’s integrated memory controller with incredibly high and unsafe voltage levels, we achieved relative stability with the XMP Profile 1. Please note, we would never condone this voltage profile for daily use, as it will most certainly harm the CPU long term. The System Agent voltage is the most important consideration when dealing with high-frequency memory overclocking.
 
OP
T

TacOp280

New Member
Joined
Dec 24, 2020
Thanks for the replies, I really do appreciate it.
Great article, I guess that settles it. Those are some crazy voltages... :screwy: and I'm on a 10700k...

After a full night of gaming the system seems very stable with XMP 1 enabled , downclocked to 4400mhz CL19-26-26 @1.5V.
4500mhz really messed with the latency for some odd reason during testing. 4400mhz testing was pretty darn good. SA/IO voltages are within safe limits mentioned in your quote.
The system still performs much better than with the previous 3600 cl17 kit so ill gladly settle with these results. And yes you're right, lowering from 4600 to 4400 or even 4000 @ 19-23-23 did not translate to much gaming wise.

I imagine I wont be the only one encountering these issues as these high frequency kits become more popular.
Thanks everyone, happy holidays!