TR Forums

Welch · Sat Mar 17, 2018 7:51 pm

https://www.newegg.com/Product/Product. ... 6820232409

Hmmmm, saweet timings. 14-14-14-34, Samsung B-Die....... Worth the cost over 16-18-18-38 Ripjaw V?

CL14 Trident Z = $214
CL16 RipJaws V = $190ish

So torn here, was really hoping RAM would have dropped by now. Need to get a kit for the wife's 1600x system.

She would probably like these too just for the prettiness - https://www.newegg.com/Product/Product. ... -_-Product
$20 off right now bring it to around the same price as the plain Jane RipJaws V

Right now both of our systems are GPU limited. Radeon 6850 and 7850 but will be upgrading cards within probably the next year, assuming the GPU world recovers from Cryptopocolypse.

DPete27 · Sat Mar 17, 2018 8:56 pm

DDR4-3000 CAS 15 for $165
Not sure where you find out if RAM is Samsung B Die or not. (It's so dumb that's a requirement for higher clocked RAM on Ryzen)

Welch · Sun Mar 18, 2018 2:03 am

So far I haven't checked for most systems I build. I use GSkill Ripjaws V and they all have not had issues running 3200 CL16 using D.O.C.P.

I was just looking at a chart of RAM speed and CL latency that supposedly shows lower latency RAM of a slower frequency as performing faster than higher frequency with loose timings. For instance, DDR4 3000 with CL14 is faster than 3200 with CL16. Not sure this is the entire picture, but had me curious.

Those low latency DIMMs I linked also have some amazingly low secondary timings. 14-14-14-34 as opposed to the usual 16-18-18-38 most other sticks are pushing. I even saw Geil sticks with CL20 @ 3200... I thought it was a typo! Yuck.

Sun Mar 18, 2018 9:03 am

Welch wrote:
I was just looking at a chart of RAM speed and CL latency that supposedly shows lower latency RAM of a slower frequency as performing faster than higher frequency with loose timings. For instance, DDR4 3000 with CL14 is faster than 3200 with CL16. Not sure this is the entire picture, but had me curious.

It's almost never that simple! :lol:

It would be more accurate to say that it might be faster, depending on workload. 14 cycles at 3000 MHz is indeed slightly less time (in absolute terms) than 16 cycles at 3200; so if you have a workload that is completely latency-bound (and does not benefit enough from the increased bandwidth of the 3200 MHz RAM to make up for the latency hit), then yeah you could see better performance with the "slower" RAM.

At the risk of (still) over-simplifying somewhat, CL is basically the overhead to initiate transfer of a memory word at a new address; additional transfers of words at sequentially increasing addresses do not incur the CL penalty. If your workload is mostly accessing single words scattered all over the place, then CL will have a large effect on performance; OTOH, if you're doing mostly larger reads/writes of blocks of words at contiguous addresses, effect of CL will be negligible, and raw clock speed will be the dominant factor in performance. Most real-world workloads will be somewhere in between.

Another way to look at it is, think of CL as being kind of like seek time on a HDD. In relative terms the hit is much smaller, but the concept is analogous in that it's the penalty you pay for not accessing the data sequentially.

Bauxite · Sun Mar 18, 2018 9:10 am

The team dark pro 14-14-14-31 was $200 with the 10% promo code, I think one of the 3200C14 gskill kits was ~190ish.

Its not just ryzen as all really fast ddr4 is samsung, the other kinds have caught up to some of the speed bands but always later in production and at slower timings.
A couple forums have decent lists, the german one is the most comprehensive I've seen so far: https://www.hardwareluxx.de/community/f ... 61530.html

Full timings don't lie, and some brands have almost no samsung chips out there at their usual resellers (lookin' at you corsair).

Chrispy_ · Sun Mar 18, 2018 10:36 am

Just calculate the actual latency in nanosenconds, not the CL rating in cycles.

Bandwidth increases only make a difference in a few edge-cases and those edge cases tend to be overblown because articles that test memory bandwidth go out of their way to find edge cases that can show the difference between different RAM timings/speeds.

In reality, low latency benefits absolutely everything, all the time.
Bandwidth only matters if there is a lack of it, in which case the difference between 3000 and 3200 is going to be negligible.

DancinJack · Sun Mar 18, 2018 10:57 am

So sad we're back to the days of trying to find tight-timed RAM.

MAKE BETTER MEMORY CONTROLLERS AMD.

thecoldanddarkone · Sun Mar 18, 2018 12:16 pm

DPete27 wrote:
DDR4-3000 CAS 15 for $165
Not sure where you find out if RAM is Samsung B Die or not. (It's so dumb that's a requirement for higher clocked RAM on Ryzen)

These are not Samsung B's. These are probably Hynix. Easiest way to get B's is to get 3000c14 or 3200c14. Also going to 3600 or faster pretty much guarantee's B's.

(no warranty expressed or given on this information lol)

Note*
I have this kit
https://www.newegg.com/Product/Product. ... -_-Product

They are single sided hynix memory modules.

Bauxite · Sun Mar 18, 2018 12:27 pm

DancinJack wrote:
So sad we're back to the days of trying to find tight-timed RAM.

MAKE BETTER MEMORY CONTROLLERS AMD.

All cpus benefit from better ram. Samsung dram is just objectively superior by a decent margin right now, intel happens to be more forgiving of the lesser stuff (and the lesser stuff aims at intel to some degree). The ryzen bios updates have shown that not all limitations are in the actual hardware design either.

The entire general computing architecture (we're up to what, 7 layers now?) is about masking latency as much as possible from the cpu cores. It has such a huge effect many designs are willing to spend significant resources precalculating multiple results (with interesting consequences) and rearranging popular code types to an entirely different architecture on the fly (x86 is nearly 100% instruction "emulation" now) just to keep them fat, happy and busy.

synthtel2 · Sun Mar 18, 2018 2:05 pm

Chrispy_ wrote:
In reality, low latency benefits absolutely everything, all the time.
Bandwidth only matters if there is a lack of it, in which case the difference between 3000 and 3200 is going to be negligible.

Rarely is there a lack of bandwidth for a program as a whole, but individual algorithms may easily be bandwidth-limited, and most programs will have something describable that way. In the extreme case, consider something like zeroing out an allocation. It probably doesn't take long, but nonetheless, the more bandwidth is handy the quicker the CPU can move on to doing other things.

Hooking more cores up to a given number of RAM channels should usually also increase the importance of bandwidth relative to latency.

I have seen noticeable boosts in practical gaming from increasing memory bandwidth at constant absolute latency. The biggest confounding factor in that seems to be how increasing RAM speed tends to loosen a lot of minor timings if not dialling everything in manually.

Bauxite wrote:
Samsung dram is just objectively superior by a decent margin right now

^ That. ^

Most of AMD's problem seems to have been weird auto timings at higher speeds, and most of that seems to be fixed now. That said, if you don't want to have to tweak anything and don't want to pay for B-die, 3000 is probably safer.

Welch · Mon Mar 19, 2018 12:53 pm

Yep, I'm aware of how RAM latencies work (albeit, a bit rusty), very good overview explanation of them. These were specs I actually cared about too much back in the day. I should have mentioned that I would be using 3200 regardless, no plans on using a slower frequency and hoping tghter timings will pickup the slack.

If things haven't changed too much, gaming or anything latency dependent used to be where tighter timings shined. I'm not sure if this is still the case, but I'd imagine it is.

So I know the whole RGB thing is a joke, but not to the wife who loves that her new B350-Gaming 3 has RGB and matches her Corsair RGB :lol:

. $190ish for 16GB or $220ish for lower latency no RGB. Is it worth the extra 30-40 to drop down to 14-14-14-34. Would love to hear some opinions to sway my purchase.

synthtel2 · Mon Mar 19, 2018 2:25 pm

What kinds of games at what kind of target framerate (assuming gaming is the primary heavy workload)?

Welch · Mon Mar 19, 2018 3:59 pm

At this point we have just been playing Planetside 2 and Path of Exile. Once the GPU market is somewhat normal and we can get with more punch we will probably start playing Ark Survival again and a few other resource intensive games. I wouldn't say we have a target frames either at this point, just as much as I can get out of it. In my case, with my Freesync monitor, I'd shoot for maintaining 50+ in most anything I play. No issues with slightly lowering the graphics. I've got the Nixeus EDG 27, so 30-144hz on the Freesync range and LFC for below that.

A game like Ark I'd imagine is going to be more CPU/RAM intensive on top of GPU. Before I was sitting around 13GB total system memory used for my entire system while running it, sometimes peaking 14gb.

MOSFET · Mon Mar 19, 2018 3:59 pm

These G.Skill modules don't work at 3200 in either Ryzen or Kaby Lake systems, for me. They are Hynix. I got 2400 Ryzen and 2800 KL. YMMV.

Welch · Mon Mar 19, 2018 4:18 pm

MOSFET wrote:
These G.Skill modules don't work at 3200 in either Ryzen or Kaby Lake systems, for me. They are Hynix. I got 2400 Ryzen and 2800 KL. YMMV.

I hate to ask the obvious but.... have you updated your BIOS? What board are you using on either of them? This seems odd and perhaps you got a bad kit. Have you tried running a single DIMM to see if it runs faster or 3200? How about using D.O.C.P instead of manually setting it?

I've setup about 12 Ryzen systems thus far and not one of them had issue with running 3200, with the exception of the very first system that has since had a BIOS revision and now works flawlessly at 3200. All of the kits except that first one were 2x4 kits, Ripjaws V 16-18-18-38 or 16-16-16-38 kits which are likely Hynix.

thecoldanddarkone · Mon Mar 19, 2018 4:55 pm

Welch wrote:
Yep, I'm aware of how RAM latencies work (albeit, a bit rusty), very good overview explanation of them. These were specs I actually cared about too much back in the day. I should have mentioned that I would be using 3200 regardless, no plans on using a slower frequency and hoping tghter timings will pickup the slack.

If things haven't changed too much, gaming or anything latency dependent used to be where tighter timings shined. I'm not sure if this is still the case, but I'd imagine it is.

So I know the whole RGB thing is a joke, but not to the wife who loves that her new B350-Gaming 3 has RGB and matches her Corsair RGB . $190ish for 16GB or $220ish for lower latency no RGB. Is it worth the extra 30-40 to drop down to 14-14-14-34. Would love to hear some opinions to sway my purchase.

If the wife likes RBG get her RBG modules. This is for your wife, who apparently likes shiny. Welch these modules like to shine (well glow vibrantly).

Welch · Mon Mar 19, 2018 6:36 pm

thecoldanddarkone wrote:
Welch wrote:
Yep, I'm aware of how RAM latencies work (albeit, a bit rusty), very good overview explanation of them. These were specs I actually cared about too much back in the day. I should have mentioned that I would be using 3200 regardless, no plans on using a slower frequency and hoping tghter timings will pickup the slack.

If things haven't changed too much, gaming or anything latency dependent used to be where tighter timings shined. I'm not sure if this is still the case, but I'd imagine it is.

So I know the whole RGB thing is a joke, but not to the wife who loves that her new B350-Gaming 3 has RGB and matches her Corsair RGB . $190ish for 16GB or $220ish for lower latency no RGB. Is it worth the extra 30-40 to drop down to 14-14-14-34. Would love to hear some opinions to sway my purchase.

If the wife likes RBG get her RBG modules. This is for your wife, who apparently likes shiny. Welch these modules like to shine (well glow vibrantly).

There is a Moana joke in their somewhere :lol:

She likes Shiny, but also likes her system to kick ass like mine. Now she is on the same CPU as me. Granted, I'm not sure she would notice the difference of the low latency in the near future lol.

Was just hoping someone had feedback on the real world difference of the high and low latency stuff, especially considering the Infinity Fabric on Ryzen.

thecoldanddarkone · Mon Mar 19, 2018 6:47 pm

There is a Moana joke in their somewhere

She likes Shiny, but also likes her system to kick ass like mine. Now she is on the same CPU as me. Granted, I'm not sure she would notice the difference of the low latency in the near future lol.

Was just hoping someone had feedback on the real world difference of the high and low latency stuff, especially considering the Infinity Fabric on Ryzen.

Hopefully someone can provide you feedback on that. Unfortunately I don't have a Ryzen system (or have had a chance to test one).

My wife only cares if the pc is able to watch video and do paperwork. If the 2500 in her pc wasn't fast enough I'd just pull out the 4930k/x79ws/32gb ram (it's been sitting in a box since December, lol).

MOSFET · Mon Mar 19, 2018 6:51 pm

Welch wrote:
I hate to ask the obvious but.... have you updated your BIOS? What board are you using on either of them? This seems odd and perhaps you got a bad kit. Have you tried running a single DIMM to see if it runs faster or 3200? How about using D.O.C.P instead of manually setting it?

-Always up to date. Asus ROG X370-F and Prime X370 Pro, which tend to run on the same firmware rev.
-DOCP tried first, and more often than manual settings. Manual settings caused all sorts of other issues on the Ryzen mobos, like dynamic clock speeds would go away, voltages would lock at maximums, just general weirdness. It's enough weirdness when overclocking that I run both Ryzen 5 1600's on Auto now.
-When CPU is on Auto settings and is loaded, it seems to run 4 cores at 3400 and 2 cores at 3700.
-The Ryzen 5 that runs Windows, where I can fire up HWMonitor, has logical cores 0 through 13, and 6 and 7 are disabled.
-If Dropbox had not taken away the Public folder, I would post a screenshot. Still learning image hosting alternatives.

My Ryzens are worthless if they aren't stable, so fretting about RAM is something I have to leave for others. That's difficult, as I'm as much of a tweaker as anyone else here, and although I refuse to be a "fan" of a company, many who are less strict with definitions would call me an AMD-lover / Intel-hater (which again, is not necessarily accurate, but that would be the perception.)

anubis44 · Mon Mar 19, 2018 9:01 pm

Yes. It's worth it. I tried CL16 2933 RAM with my Ryzen 7 1700, and you can feel the difference with DDR4 CL14 3200MHz ram. Just do it.

synthtel2 · Mon Mar 19, 2018 11:36 pm

Welch wrote:
At this point we have just been playing Planetside 2 and Path of Exile. Once the GPU market is somewhat normal and we can get with more punch we will probably start playing Ark Survival again and a few other resource intensive games. I wouldn't say we have a target frames either at this point, just as much as I can get out of it. In my case, with my Freesync monitor, I'd shoot for maintaining 50+ in most anything I play. No issues with slightly lowering the graphics. I've got the Nixeus EDG 27, so 30-144hz on the Freesync range and LFC for below that.

The B-die stuff will definitely be worth a boost in Planetside (R7 1700 / EDG 27 owner here who has spent way too much time trying to get every last frame out of that game). CPU performance has been degrading again and really hates AMD; 50 in the biggest fights is a bit of a tall order, but with 3200C14 and shadows turned down it should be close.

Welch · Tue Mar 20, 2018 5:24 am

MOSFET wrote:
Welch wrote:
I hate to ask the obvious but.... have you updated your BIOS? What board are you using on either of them? This seems odd and perhaps you got a bad kit. Have you tried running a single DIMM to see if it runs faster or 3200? How about using D.O.C.P instead of manually setting it?

-Always up to date. Asus ROG X370-F and Prime X370 Pro, which tend to run on the same firmware rev.
-DOCP tried first, and more often than manual settings. Manual settings caused all sorts of other issues on the Ryzen mobos, like dynamic clock speeds would go away, voltages would lock at maximums, just general weirdness. It's enough weirdness when overclocking that I run both Ryzen 5 1600's on Auto now.
-When CPU is on Auto settings and is loaded, it seems to run 4 cores at 3400 and 2 cores at 3700.
-The Ryzen 5 that runs Windows, where I can fire up HWMonitor, has logical cores 0 through 13, and 6 and 7 are disabled.
-If Dropbox had not taken away the Public folder, I would post a screenshot. Still learning image hosting alternatives.

My Ryzens are worthless if they aren't stable, so fretting about RAM is something I have to leave for others. That's difficult, as I'm as much of a tweaker as anyone else here, and although I refuse to be a "fan" of a company, many who are less strict with definitions would call me an AMD-lover / Intel-hater (which again, is not necessarily accurate, but that would be the perception.)

I'm trying to figure what the odds are that you would have issues with BOTH boards, I'd guess not likely. I'd suggest you have something wrong with the RAM then. Have you tried contacting G.Skill? They have a lifetime warranty on the RAM and if it doesn't run it's rated speeds in two completely different platforms, then I'd venture to say something isn't right, especially when it is off by a good amount, not just 3000 running as 2933.

This guys video about his Ryzen system is especially a good example of how the board could be at fault though, at least if it isn't the DIMMs.
https://youtu.be/k1GZo3WBOKo?t=4m59s

As far as cores being disabled!!? Have you never been able to get all cores active? If so, then again.. something in your system isn't right and I'd be looking to determine what all is wrong with your hardware. Whether that is a motherboard (or two), the CPU and subsequently memory controller? I'm just not sure, you have too many strange things going on and if I were in your boat, I'd for sure not accept that with as much cash as you probably have wrapped up in those systems. Like I said before, I've got about 12 Ryzen builds under my belt now, and not a single one won't boot into 3200 on Hynix RipJaws V based memory. It would be another story MAYBE if you were using 4 DIMMs.

synthtel2 wrote:
The B-die stuff will definitely be worth a boost in Planetside (R7 1700 / EDG 27 owner here who has spent way too much time trying to get every last frame out of that game). CPU performance has been degrading again and really hates AMD; 50 in the biggest fights is a bit of a tall order, but with 3200C14 and shadows turned down it should be close.

With your EDG 27, did you ever notice any flicker in PS2? I loaded PS2 up on my temporary secondary rig (an old A8-3850) with an RX 550 and wanted to see what Freesync was all about as I'm still on that 7850 which doesn't do Freesync unfortunately. My frames were high enough in warpgate sitting easily in the 50-70 range, everything set to low.... but I was getting a sort of strobe effect from the monitor and noticed the refresh was jumping and hitting below 30 randomly. I chalked it up to either PS2 not being compatible or the fact that I grabbed a random DisplayPort cable that I hadn't confirmed to be 1.4a. However, in Windows it runs 144hz just fine and if I disabled Freesync, the game ran fine with the 144hz. Also I ran Freesync enabled while running Path of Exile as a test and didn't notice any strobing. Was very confused but haven't had enough time to look into it.

synthtel2 · Tue Mar 20, 2018 3:33 pm

I'm stuck on a GTX 960 at the moment, so can't Freesync to test it.

I do have one idea - if smoothing isn't turned on (checkbox in graphics settings), certain arrangements of physics objects in the game can cause some very heavy microstutter (alternating slow/fast frames) for nearby players. I've seen Galaxies getting stuck in warpgate geometry cause this before. This stutter could easily be heavy enough to confuse Freesync.

Smoothing is generally a good option to have turned on in PS2 anyway when it capping your framerate at your monitor's refresh rate isn't a problem. It 90% nullifies both a stupid source of input lag which shows up when GPU-bound and PS2's rate of fire dependency on framerate. The alternative for those fixes is lowering MaximumFPS in UserOptions.ini (until you're bound by the config value instead of your GPU and until your net framerate is rarely less than half the config value, respectively).

synthtel2 · Fri Mar 23, 2018 12:41 am

synthtel2 wrote:
CPU performance has been degrading again and really hates AMD; 50 in the biggest fights is a bit of a tall order, but with 3200C14 and shadows turned down it should be close.

Disregard that, turns out model quality is the trick (animation LOD distances). Of course none of my testing caught that (until now) because I mostly wasn't testing with other players around, because they make test results inconsistent. :oops:

After turning that down to low, my framerate in big fights has about doubled, and now hardly ever goes below 60 (Zen at 3.25 GHz / 2666CL14, core count doesn't matter due to PS2 being only four-wide).

Welch · Fri Mar 23, 2018 2:22 am

synthtel2 wrote:
synthtel2 wrote:
CPU performance has been degrading again and really hates AMD; 50 in the biggest fights is a bit of a tall order, but with 3200C14 and shadows turned down it should be close.

Disregard that, turns out model quality is the trick (animation LOD distances). Of course none of my testing caught that (until now) because I mostly wasn't testing with other players around, because they make test results inconsistent.

After turning that down to low, my framerate in big fights has about doubled, and now hardly ever goes below 60 (Zen at 3.25 GHz / 2666CL14, core count doesn't matter due to PS2 being only four-wide).

Surprised that is all you are getting, your GPU is much more powerful than my 7850 2GB... What sort of res are you pushing out? I've been playing PS2 for roughly 4 years, sad to say almost exclusively. I just can't get into games because of time constraints with work and kids. I've done a lot of fine tuning to my PS2 over the years :lol:

, I've got it running damn fine. The biggest jump was going from my i5-2500k to the 1600x as it simply smoothed out the jumps in frames. My rule of thumb for editing any game settings, native res is a must, lower or turn off any particle or shadows, especially when they don't change the gameplay or add anything useful. In most cases editing render distance is also a must, but sadly it seems that PS2 doesn't benefit from lowering it. So I do leave it on at least 4k if not maxing it out at least on my 7850.

I hate to say "I wish I bought that RAM" as it was $213, it's not $228

. Apparently the RGB loose timing RAM doesn't ship here for free with my ShopRunner sub, so it would cost me around $20 to ship it because newegg shipping sucks up to AK.

biffzinker · Fri Mar 23, 2018 3:43 am

Welch wrote:
[so it would cost me around $20 to ship it because newegg shipping sucks up to AK.

Also why I ordered my Phanteks case from Amazon instead of Newegg.

Welch · Fri Mar 23, 2018 5:04 am

Well it looks like a bunch of other (slower) DIMMS went on sale...

https://www.newegg.com/Product/Product. ... -_-Product

172.99 after $20 and it is ShopRunner eligible, so I can ship for free... hmmm.

I almost forgot you were in AK too Biff! Geeez, when the TR BBQ goes on, we seriously need to have a satellite fire pit going or something. I wonder how many other gerbils are from the great state? Let me know when you are down in my neck of the woods.

I've been buying all of my cases from Amazon for the last 3-4 years. I think I was single handedly the reason ShopRunner stopped supporting computer cases :lol:

. I was ordering Corsair 600t and other full tower cases like there was no tomorrow. I saw the shipping on some of those cases if you did the normal 3-5 day type shipping was about $80, I can't imagine what it was costing for 2nd day!

Airmantharp · Fri Mar 23, 2018 11:40 am

It's your wife- it's worth it to buy her what she wants, and it's not like it's significantly more expensive in terms of overall build cost

.

G.SKILL 3200C14 RGB for US$249.

synthtel2 · Fri Mar 23, 2018 5:35 pm

Welch wrote:
Surprised that is all you are getting, your GPU is much more powerful than my 7850 2GB... What sort of res are you pushing out? I've been playing PS2 for roughly 4 years, sad to say almost exclusively. I just can't get into games because of time constraints with work and kids. I've done a lot of fine tuning to my PS2 over the years , I've got it running damn fine. The biggest jump was going from my i5-2500k to the 1600x as it simply smoothed out the jumps in frames. My rule of thumb for editing any game settings, native res is a must, lower or turn off any particle or shadows, especially when they don't change the gameplay or add anything useful. In most cases editing render distance is also a must, but sadly it seems that PS2 doesn't benefit from lowering it. So I do leave it on at least 4k if not maxing it out at least on my 7850.

[Display]
FullscreenWidth=2560
FullscreenHeight=1440
RenderQuality=1.000000

[Rendering]
GraphicsQuality=1
TextureQuality=0
ShadowQuality=1
LightingQuality=1
EffectsQuality=1
TerrainQuality=1
ModelQuality=1
RenderDistance=6000.000000
VerticalFOV=68
ParticleLOD=0
ParticleDistanceScale=0.650000
FogShadowsEnable=0
VSync=0
AO=0
MaximumFPS=150
UseLod0a=0
BloomEnabled=0
Smoothing=1

[Terrain]
RenderFlora=Off

TextureQuality=0 corresponds to ultra, shadows on low or medium have reasonable CPU load and can be a useful situational awareness thing, and there's no way to set up MaxFPS and smoothing that's truly optimal for all three of input lag, rate of fire, and framerate. I seem to have ended up changing smoothing and RenderDistance depending on what's going on in-game. Render distance can help out a bit, but mostly in the 500-1500 range. I set it to 500 for dense CQC stuff and back to 6000 if sniping, flying, etc. I'd often turn RenderQuality (resolution) down a bit too, more for its indirect input lag benefits than its framerate benefits, but running native res really does make it easier to see cloaked infiltrators.

The result is GPU-bound around 90 in higher-detail scenes and 120+ in more moderate ones, CPU-bound around 60 in very dense fights and 100+ in more moderate ones. That extra 450 MHz you've got on your CPU is no small deal, and I am usually CPU-limited. 4C4T to 4C8T/6C6T/+ is a big boost, as it caps at about 80% utilization of whatever it's on and does have occasional jobs that go beyond 4T wide, but 6C12T to 8C16T is worth nil. On the GPU side of things, it also seems to like AMD a bit better than Nvidia, per otherwise equivalent cards.

TR Forums

DDR4 3200 - CL14 - To be or not to be

DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Re: DDR4 3200 - CL14 - To be or not to be

Who is online