Ballistix Sport LT Overclocking

[DeliciousStorage]

@MegaPinches these are the slightly updated ones which I think should work with most E-die kits & CPUs, should get you ~66ns in AIDA64 on a 2x SR kit:
tCL 16, tRCD(RD) 20, tRCD(WR) 20, tRP 16, tRAS 38, tRC 58
tRRDS 4, tRRDL, 6, tFAW 16, tWTRS 4, tWTRL 12, tWR 16, tRDRD SCL 4, tWRWR SCL 4, tRFC 580
tCWL 16, tRTP 12, tRDWR 9, tWRRD 2, tWRWR SC 1, tWRWR SD 7, tWRWR DD 7, tRDRD SC 1, tRDRD SD 5, tRDRD DD 5, tCKE 1 (or leave tCKE auto)
{tPage Auto, tRFC2 Auto, tRFC4 Auto}
ProcODT 40 Ohm, CMD2T 1T, GearDownMode Enabled, PowerDownMode Disabled
RTT, CAD Bus settings all on Auto
Voltages: 1.45v vDIMM (VTT should be half vDIMM, but on auto it's set to half anyway), 1.1v vSOC, 1.05v cLDO_VDDG, 1v cLDO_VDDP

If stuttering, up vSoC to 1.125v and VDDG to 1.075v (that fixed it for me). Loosened some tertiaries just to be on the safe side and from what I noticed lowering a few timings too much can actually have a negative impact on latency.

Thanks, these settings worked well for me. I lowered tRFC to 560 and the voltages to 1.4v vDIMM, 0.95v cLDO_VDDG, 0.9v cLDO_VDDP. Seems very stable, passed an overnight Karhu RAMTest and a few hours TM5 with extreme config. I ran it for a while without changing the VDDG and VDDP voltages but realized that they were set to 1.15v and 1.1v when I can across this post, hopefully it didn't cause any harm. I tried setting vDIMM to 1.35v but it failed Karhu after 1 minute and TM5 started spitting out errors almost immediately.

Here is my benchmark

 

[DeliciousStorage]

I've read through some more of this thread and I'm unsure of a few things, this is the first time I've overclocked my RAM. Some comments mentioned game stutters and crackling audio when the FCLK is too high (1866mhz). I haven't tried gaming with it yet but it has passed overnight Karhu and TM5 RAM tests with 0 errors, is it still possible I could have these stuttering and crackling issues? If so, what are the technical reasons why this happens if there are no errors? It makes me wonder if it could cause other weird problems that I haven't noticed yet and so it might be better to go down to 1800 FCLK.

 

[pc-guy]

[QUOTE="DeliciousStorage, post: 33378885, member: 219435"

Here is my benchmark
[/QUOTE]
Interesting results. My crucial is running at 3200MHz with the timing obtained from DRAM calculator. The latency is 70ns but the bandwidth is much higher. 58Gb/s in that reading field but it might have something to do with my ranks are dual ranked.

 

[gupsterg]

Wow. Stock Voltage. , need to drop my Volts then.
They certainly give these 8pack Teamgroup Samsung b-die “Binned” Sticks a run for there money.
Glad I went on the advice of @tamzzy when picking Ram.

Done slight tinker, AIDA64 & stability testing of ~17hrs in this ZIP.

I have now also dropped tRFC to 551, coming upto ~24hrs mixed testing, will share soon .

Here is my benchmark

Interesting results. My crucial is running at 3200MHz with the timing obtained from DRAM calculator. The latency is 70ns but the bandwidth is much higher. 58Gb/s in that reading field but it might have something to do with my ranks are dual ranked.

I have dual ranked/sided dimms and don't see bandwidth of 58Gb/s at 3200MHz, mind sharing a screenie of AIDA64, timings, etc, cheers .

 

[pc-guy]

Done slight tinker, AIDA64 & stability testing of ~17hrs in this ZIP.

I have now also dropped tRFC to 551, coming upto ~24hrs mixed testing, will share soon .



I have dual ranked/sided dimms and don't see bandwidth of 58Gb/s at 3200MHz, mind sharing a screenie of AIDA64, timings, etc, cheers .

nope. Just checked Aida. It’s 49800MB/S. Pulling a system together at 2am and running stress testing makes one hallucinate

time to get those clocks up.

 

[gupsterg]

3800C16R9.3

Spoiler

DRAM CAS# Latency [16]
DRAM RAS# to CAS# Read Delay [19]
DRAM RAS# to CAS# Write Delay [19]
DRAM RAS# PRE Time [17]
DRAM RAS# ACT Time [43]
Trc [60]
TrrdS [6]
TrrdL [8]
Tfaw [32]
TwtrS [4]
TwtrL [12]
Twr [16]
TrdrdScl [4]
TwrwrScl [4]
Trfc [551]
Trtp [8]

3xAIDA64

~15K% PASS in Kahru RAM Test

nope. Just checked Aida. It’s 49800MB/S. Pulling a system together at 2am and running stress testing makes one hallucinate

time to get those clocks up.

Ahh, OK .

*** edit ***

ZIP Link, order files by time, ~49hrs back to back mixed testing of 3800C16R9.3.

I have owned Zen/Zen+/Zen2, on AM4 & sTR4, used several kits of G.Skill, all Samsung B die, bins 3200C14, 3600C15, 4000C18. None of those have given me the OC buzz that the Crucial Ballistix Sport LT 2x16GB kit (BLS2K16G4D32AESB) have . Phenomenal value for sure these kits.

 

[Mikkinen]

Hello, after many tests with 2700X and BLS16G4D30AESC.M16FE ram, in my opinion these are the two best results.
At 3549 cl16 the ram was stable, but unstable on the whole (ibt avx does not pass, same thing with ram b-die 2x8gb, I presume my cpu limit).
I think I keep the ram at 3200 cl 14 at 1.35v, it doesn't seem to make sense for the game considered 3481 cl14 at 1.45v.
Thanks for the advice I received

 

[Scougar]

So many good threads...this is the exact thread I need.

 

[gupsterg]

3800C16R9.4

Spoiler: Timings

DRAM CAS# Latency [16]
DRAM RAS# to CAS# Read Delay [19]
DRAM RAS# to CAS# Write Delay [19]
DRAM RAS# PRE Time [17]
DRAM RAS# ACT Time [43]
Trc [60]
TrrdS [6]
TrrdL [8]
Tfaw [32]
TwtrS [4]
TwtrL [12]
Twr [16]
TrdrdScl [4]
TwrwrScl [4]
Trfc [532]
Trtp [8]

~24hrs back to back mixed testing ZIP.

 

[Mikkinen]

Which of you who went from zen + to zen2 what major differences have you noticed? Can higher frequencies and better timings be obtained with less voltage on the ram?

 

[gupsterg]

Which of you who went from zen + to zen2 what major differences have you noticed

I've had Zen, Zen+ and Zen2. Zen2 is the most refined for boost and other aspects. For example SW monitoring has more info than other gens. Ryzen Master has more settings and information available on Zen2.

Can higher frequencies and better timings be obtained with less voltage on the ram?

I have used same Samsung B die kit with Zen, Zen+ and Zen2. Let's say I set VDIMM at ~1.35V, Zen would top out ~3333MHz for me, Zen+ ~3533MHz and Zen2 is ~3800MHz, FCLK/UCLK/MEMCLK 1:1:1.

 

[Mikkinen]

I have used same Samsung B die kit with Zen, Zen+ and Zen2. Let's say I set VDIMM at ~1.35V, Zen would top out ~3333MHz for me, Zen+ ~3533MHz and Zen2 is ~3800MHz, FCLK/UCLK/MEMCLK 1:1:1.

Ok, so does the cpu capacity also affect the voltage needed to reach a certain frequency? For example with b-die 2x8gb I got 3481cl14 with 1.41v, even if not comparable with ballistix because 2x16gb dr.

 

[gupsterg]

Ok, so does the cpu capacity also affect the voltage needed to reach a certain frequency? For example with b-die 2x8gb I got 3481cl14 with 1.41v, even if not comparable with ballistix because 2x16gb dr.

CPU IMC and RAM IC quality would be factors dictating what VDIMM you may need, besides timings used. I'd start a separate thread or post in Samsung B die related thread or we'd be taking this one off track.

 

[Troezar]

After some tinkering got some optimized timings for 3733Mhz C16 for E-die that should work on most E-die kits I feel:
tCL 16, tRCD(RD) 20, tRCD(WR) 20, tRP 16, tRAS 38, tRC 58
tRRDS 4, tRRDL, 6, tFAW 20, tWTRS 4, tWTRL 12, tWR 16, tRDRD SCL 4, tWRWR SCL 4, tRFC 560
tCWL 16, tRTP 12, tRDWR 8, tWRRD 2, tWRWR SC 1, tWRWR SD 6, tWRWR DD 6, tRDRD SC 1, tRDRD SD 4, tRDRD DD 4, tCKE 1 (or leave tCKE auto)
{tPage Auto, tRFC2 Auto, tRFC4 Auto}
ProcODT 40 Ohm, CMD2T 1T, GearDownMode Enabled, PowerDownMode Disabled
RTT, CAD Bus settings all on Auto
Voltages: 1.45v vDIMM (VTT should be half vDIMM, but on auto it's set to half anyway), 1.1v vSOC, 1.05v cLDO_VDDG, 1v cLDO_VDDP
Some of the voltages maybe be a bit generous, but if the above is stable with some extensive testing, then you can always try lower the voltages gradually (they're all still within safe ranges for everything though). Can also try tightening tFAW to 16 and tWR to 14 or 12.

They should give pretty good speeds for read/write/copy in AIDA, with latency of about ~68ns for 2x SR DIMMs or 70~71ns for 4x SR DIMMs.

@AndreiD Thanks, I tried these at 3600MHz and have 32GB (2x16) of BLS16G4D32AESB.M16

Seems to be working ok so far at 1.45V, need to test it to see how it is performing. I was trying C14 which worked but huge number of errors, probably a tad ambitious!

 

[r22snapper]

As above. I know nothing about memory overclocking, put AndreiD's settings in and have 32GB running at 3733 with 67.4ns latency. Thanks very much!

 

[r22snapper]

Got a couple of memtest errors at 1.45, dropped to 1.42 and flawless, thanks again Andrei.

 

[Robertsanders]

how can i post my OC results? i been getting 60-62ns r5 2600x

 

[Kucuboy]

Hi, I am new to the forum, cheers. I have a B450 tomahawk Max paired with a 2700X and Ballistix Sport LT 16GB (8x2) 3000Mhz CL15 (BLS2K8G4D30AESEK). My mobo bios (latest 350 version) does not show VDDG Voltage. But I notice in some of the screenshots of others here that it can be adjusted in Ryzen Master, but I cant to find that screen at all. I've attached my screenshots I hope you could point me to the right direction. Tq

 

[Scougar]

Has anyone been getting kernel power (41) errors when upping frequency on just the memory?

Trying to run at 3600 and getting this.

1.4v mem. 1.11v soc. (1.3v on cpu core).

 

[roadie]

I need to get into optimising my 3800X and 16Gb 3000Mhz kit! If I could get a nice and stable 3600/1800Mhz on the memory and IF I'd be happy.