F5F Stay Refreshed Hardware Desktop What's the difference between TFRC and TRC? (AMD versus Intel in the 5th generation timing)

What's the difference between TFRC and TRC? (AMD versus Intel in the 5th generation timing)

What's the difference between TFRC and TRC? (AMD versus Intel in the 5th generation timing)

Pages (3): 1 2 3 Next
L
LOL0
Member
191
01-19-2016, 07:34 AM
#1
AMD: The RIGHT value should equal tRP plus tRAS, which gives the correct tRC. Intel’s tRFC of 467 seems off—why not align it properly? Minimum safe tRFC calculation is important; reducing from 467 to 50 might be too drastic.
L
LOL0
01-19-2016, 07:34 AM #1

AMD: The RIGHT value should equal tRP plus tRAS, which gives the correct tRC. Intel’s tRFC of 467 seems off—why not align it properly? Minimum safe tRFC calculation is important; reducing from 467 to 50 might be too drastic.

H
Hermo
Junior Member
21
01-19-2016, 07:34 AM
#2
tRC and tRFC are entirely distinct, with each platform incorporating those values. CPU-Z isn't displaying them properly for some reason. tRFC shouldn't register around 50 cycles, let alone Samsung B-Die, which struggles to go below roughly 250 cycles. Overall, everything appears normal for both systems.
H
Hermo
01-19-2016, 07:34 AM #2

tRC and tRFC are entirely distinct, with each platform incorporating those values. CPU-Z isn't displaying them properly for some reason. tRFC shouldn't register around 50 cycles, let alone Samsung B-Die, which struggles to go below roughly 250 cycles. Overall, everything appears normal for both systems.

J
JdGamingShow
Member
180
01-19-2016, 07:34 AM
#3
They are totally different, which means altering it like in the Ryzen situation would be risky. For Ryzen the calculation should be 19 plus 39 equals 58, so what method do you use for Intel? Is the default setting best?
J
JdGamingShow
01-19-2016, 07:34 AM #3

They are totally different, which means altering it like in the Ryzen situation would be risky. For Ryzen the calculation should be 19 plus 39 equals 58, so what method do you use for Intel? Is the default setting best?

B
BenTGreat
Senior Member
437
01-19-2016, 07:34 AM
#4
Calculating an optimal tRFC isn't straightforward because each memory IC behaves differently. For instance, a B-Die might work well below 300 MHz, while a Hynix MFR struggles even at the same speeds. I recommend sticking with the default settings, even if they're a bit conservative. You probably won’t notice big performance improvements since you're already running at 2666MHz on Intel. Adjusting tRFC could impact stability significantly. If you're testing for stability, lowering it might help. I’m not very familiar with tRC or tRFC—this isn’t the same as tRFC, and it’s part of the DDR4 standard. Both platforms have their own timing values, and tools like CPU-Z don’t display all DDR4 details.
B
BenTGreat
01-19-2016, 07:34 AM #4

Calculating an optimal tRFC isn't straightforward because each memory IC behaves differently. For instance, a B-Die might work well below 300 MHz, while a Hynix MFR struggles even at the same speeds. I recommend sticking with the default settings, even if they're a bit conservative. You probably won’t notice big performance improvements since you're already running at 2666MHz on Intel. Adjusting tRFC could impact stability significantly. If you're testing for stability, lowering it might help. I’m not very familiar with tRC or tRFC—this isn’t the same as tRFC, and it’s part of the DDR4 standard. Both platforms have their own timing values, and tools like CPU-Z don’t display all DDR4 details.

S
SkelitonMage
Member
51
01-19-2016, 07:34 AM
#5
I won't chase peak performance, the first RAM I got was from SK Hynix and it handled up to 3600MHz CL18-19-19-39 fine. On my AMD rig it worked smoothly at 2666MHz CL16-18-18-39. The other one seems like a dead end B-die, probably Samsung 3200MHz CL16-18-18-38. I don’t have much confidence in these, so looking at the weakest part—Hynix or B-die?—I think I should go with the less powerful ones and stick to the minimum specs. X-X
S
SkelitonMage
01-19-2016, 07:34 AM #5

I won't chase peak performance, the first RAM I got was from SK Hynix and it handled up to 3600MHz CL18-19-19-39 fine. On my AMD rig it worked smoothly at 2666MHz CL16-18-18-39. The other one seems like a dead end B-die, probably Samsung 3200MHz CL16-18-18-38. I don’t have much confidence in these, so looking at the weakest part—Hynix or B-die?—I think I should go with the less powerful ones and stick to the minimum specs. X-X

H
HudsonNZ
Member
63
01-19-2016, 07:34 AM
#6
Is the B-Die in the Intel PC? From what I understand, you can't exceed 2666Mhz because you're using a B460 chip and a locked processor that restricts memory overclocking. It seems optimizing the timings on the Intel platform isn't necessary since your main limitation is frequency. However, since it's a B-Die design, you might achieve very low clock speeds at 2666Mhz.
H
HudsonNZ
01-19-2016, 07:34 AM #6

Is the B-Die in the Intel PC? From what I understand, you can't exceed 2666Mhz because you're using a B460 chip and a locked processor that restricts memory overclocking. It seems optimizing the timings on the Intel platform isn't necessary since your main limitation is frequency. However, since it's a B-Die design, you might achieve very low clock speeds at 2666Mhz.

O
opticgunship
Posting Freak
815
01-19-2016, 07:34 AM
#7
It's actually pretty funny at 1.35v—both sets managed only 13-15-15, which isn't great at all...
O
opticgunship
01-19-2016, 07:34 AM #7

It's actually pretty funny at 1.35v—both sets managed only 13-15-15, which isn't great at all...

9
992x
Senior Member
506
01-19-2016, 07:34 AM
#8
It seems odd even for the simplest B-Die, are you certain it's a B-Die? You might want to reduce the CL separately and keep the rest at auto, then test at 1.35V.
9
992x
01-19-2016, 07:34 AM #8

It seems odd even for the simplest B-Die, are you certain it's a B-Die? You might want to reduce the CL separately and keep the rest at auto, then test at 1.35V.

T
Tavado
Senior Member
505
01-19-2016, 07:34 AM
#9
It's 100% a Samsung, possibly a B-Die? Buildzoid mentioned in a video that 3200MHz CL16 models from Samsung are frequently rejected or considered low-quality tiers, so it might be that.
T
Tavado
01-19-2016, 07:34 AM #9

It's 100% a Samsung, possibly a B-Die? Buildzoid mentioned in a video that 3200MHz CL16 models from Samsung are frequently rejected or considered low-quality tiers, so it might be that.

T
Tralalapoo
Junior Member
45
01-19-2016, 07:34 AM
#10
That's 2666Mhz? I thought the Hynix board I'm using is more likely a Hynix CJR, not B-Die.
T
Tralalapoo
01-19-2016, 07:34 AM #10

That's 2666Mhz? I thought the Hynix board I'm using is more likely a Hynix CJR, not B-Die.

Pages (3): 1 2 3 Next