F5F Stay Refreshed Power Users Overclocking Adjust the BIOS settings for your AMD 7 2700X and ROG Strix X470-F gaming graphics card.

Adjust the BIOS settings for your AMD 7 2700X and ROG Strix X470-F gaming graphics card.

Adjust the BIOS settings for your AMD 7 2700X and ROG Strix X470-F gaming graphics card.

Pages (3): Previous 1 2 3
G
gogofrgl1234
Senior Member
718
03-01-2020, 08:46 AM
#21
You've already adjusted your bios, so the positive voltage offset shouldn't be an issue. The first bios I handled, probably the 4205, had voltages configured at a very high level. On AUTO with stock settings, I observed voltage spikes of 1.55V, significantly higher than what I previously experienced at 4.35Ghz and my current positive offset. Each subsequent bios update from Asus improved the "stock" voltages, but achieving an overclocked state now requires a positive offset for stability. The voltages I'm currently monitoring are actually lower than those seen with the initial bios and AUTO settings.
G
gogofrgl1234
03-01-2020, 08:46 AM #21

You've already adjusted your bios, so the positive voltage offset shouldn't be an issue. The first bios I handled, probably the 4205, had voltages configured at a very high level. On AUTO with stock settings, I observed voltage spikes of 1.55V, significantly higher than what I previously experienced at 4.35Ghz and my current positive offset. Each subsequent bios update from Asus improved the "stock" voltages, but achieving an overclocked state now requires a positive offset for stability. The voltages I'm currently monitoring are actually lower than those seen with the initial bios and AUTO settings.

T
T5lyn
Junior Member
33
03-06-2020, 09:43 AM
#22
High performance boards in the X570 lineup should pose no challenge. Expensive boards will become a concern. I intend to switch to the R7 3800X or the R9 3900X after reviewing independent reviews, but I’ll remain with my Strix X470-F. The Zen 2 generation boards are expected to be costly, featuring cutting-edge technology not previously used by any manufacturer (such as PCIe 4.0 and numerous high-speed features for M.2 drives and GPUs). These newer boards come equipped with large heat sinks and integrated cooling fans. High-end models boast extremely powerful VRMs, with the highest I’ve heard so far being 14 phases. Given the advanced technology, significant power consumption, and demanding cooling requirements, these boards will likely remain expensive and may have unresolved issues during initial testing. I don’t believe PCIe 4.0 justifies the premium unless high-end GPUs capable of using it are available within a year or so. I might consider upgrading when the X670 and Zen 2+ models arrive, or wait for the X770 and 5nm Zen 3. By then, they should have resolved PCIe 4.0 challenges, run cooler, be more reliable, and probably cost less. Many premium boards seem tailored for the R9 16 core, but a 12–14 phase VRM might be excessive even for 12 cores with 24 threads. For my content creation work, a 12-core setup would be appealing, but I’d also prefer a more efficient, higher-clocked 8-core model offering about a 15% improvement in IPC.
T
T5lyn
03-06-2020, 09:43 AM #22

High performance boards in the X570 lineup should pose no challenge. Expensive boards will become a concern. I intend to switch to the R7 3800X or the R9 3900X after reviewing independent reviews, but I’ll remain with my Strix X470-F. The Zen 2 generation boards are expected to be costly, featuring cutting-edge technology not previously used by any manufacturer (such as PCIe 4.0 and numerous high-speed features for M.2 drives and GPUs). These newer boards come equipped with large heat sinks and integrated cooling fans. High-end models boast extremely powerful VRMs, with the highest I’ve heard so far being 14 phases. Given the advanced technology, significant power consumption, and demanding cooling requirements, these boards will likely remain expensive and may have unresolved issues during initial testing. I don’t believe PCIe 4.0 justifies the premium unless high-end GPUs capable of using it are available within a year or so. I might consider upgrading when the X670 and Zen 2+ models arrive, or wait for the X770 and 5nm Zen 3. By then, they should have resolved PCIe 4.0 challenges, run cooler, be more reliable, and probably cost less. Many premium boards seem tailored for the R9 16 core, but a 12–14 phase VRM might be excessive even for 12 cores with 24 threads. For my content creation work, a 12-core setup would be appealing, but I’d also prefer a more efficient, higher-clocked 8-core model offering about a 15% improvement in IPC.

V
VIXXStarlight1
Junior Member
5
03-06-2020, 02:26 PM
#23
In short, I'm at about 1.5 V on every core when idle (no offset in BIOS), but they fluctuate a bit. HWM indicates values above 1.4, so the average is likely around 1.47. Without the average column it's tough to judge... I haven't run the stability test with Aida64 yet.
V
VIXXStarlight1
03-06-2020, 02:26 PM #23

In short, I'm at about 1.5 V on every core when idle (no offset in BIOS), but they fluctuate a bit. HWM indicates values above 1.4, so the average is likely around 1.47. Without the average column it's tough to judge... I haven't run the stability test with Aida64 yet.

I
ItsAge
Member
229
03-06-2020, 10:55 PM
#24
I don't rely on HWmonitor for Ryzen; in my experience it wasn't showing the right data. Check the voltage readings in Ryzen Master and HWinfo64. Ryzen Master remains the reliable tool for temperature tracking. I've observed that when only a single core or four cores are under load, the voltage tends to be higher than when all cores are active. For instance, with just one or up to four cores stressed, voltages usually sit around 1.43V, whereas during full core engagement—such as in a Cinebench benchmark—the reading stabilizes near 1.36V. This isn't unexpected since the system is designed to increase voltage when boosting up to four cores for stability in its "boost mode." When overclocking via PBO, it elevates all-core settings to a higher frequency and at 4.35Ghz, the processor only needs 1.36V, which is configured in BIOS through an offset. Using an offset helps because it allows the voltage to drop to a resting level of 0.831V at 2.19Ghz when "balanced mode" is enabled in Windows. If you're overclocking with a single core, one of your cores might hit 4.35Ghz while others stay at 2.19Ghz, resulting in a voltage of 1.43V due to the processor compensating for the single-core boost. This behavior is typical.

Now, if you're seeing 1.5V reported by HWinfo64 during a Cinebench render, it suggests your system is delivering more power than necessary by default. You should start by adjusting one setting at a time—run a Cinebench test with HWinfo64 open and observe the Vcore across all cores. Aim for around 1.36V when all cores are active, which should remain stable at 4.35Ghz. If this works, you're good; otherwise, gradually lower the Vcore until stability returns, then revert to the last stable negative offset. This process helps identify the optimal setting.

Every processor and motherboard behaves differently due to variations in manufacturing. With BIOS 4205, a negative offset was necessary, while newer BIOS versions often require a positive offset. However, this depends on your specific hardware. For my setup, I needed a negative offset with BIOS 4205, but with the latest BIOS you might need a positive one—though that's still possible with your model. Once you find the right Vcore, expect voltage fluctuations between 1.44V and 1.45V, with single-core boosts possibly matching those levels. Ryzen handles spikes up to 1.55V and can sustain around 1.5V, but optimizing it should keep voltages within a safe range—no more than 1.46V in boost mode, with occasional brief spikes of 1.5V.
I
ItsAge
03-06-2020, 10:55 PM #24

I don't rely on HWmonitor for Ryzen; in my experience it wasn't showing the right data. Check the voltage readings in Ryzen Master and HWinfo64. Ryzen Master remains the reliable tool for temperature tracking. I've observed that when only a single core or four cores are under load, the voltage tends to be higher than when all cores are active. For instance, with just one or up to four cores stressed, voltages usually sit around 1.43V, whereas during full core engagement—such as in a Cinebench benchmark—the reading stabilizes near 1.36V. This isn't unexpected since the system is designed to increase voltage when boosting up to four cores for stability in its "boost mode." When overclocking via PBO, it elevates all-core settings to a higher frequency and at 4.35Ghz, the processor only needs 1.36V, which is configured in BIOS through an offset. Using an offset helps because it allows the voltage to drop to a resting level of 0.831V at 2.19Ghz when "balanced mode" is enabled in Windows. If you're overclocking with a single core, one of your cores might hit 4.35Ghz while others stay at 2.19Ghz, resulting in a voltage of 1.43V due to the processor compensating for the single-core boost. This behavior is typical.

Now, if you're seeing 1.5V reported by HWinfo64 during a Cinebench render, it suggests your system is delivering more power than necessary by default. You should start by adjusting one setting at a time—run a Cinebench test with HWinfo64 open and observe the Vcore across all cores. Aim for around 1.36V when all cores are active, which should remain stable at 4.35Ghz. If this works, you're good; otherwise, gradually lower the Vcore until stability returns, then revert to the last stable negative offset. This process helps identify the optimal setting.

Every processor and motherboard behaves differently due to variations in manufacturing. With BIOS 4205, a negative offset was necessary, while newer BIOS versions often require a positive offset. However, this depends on your specific hardware. For my setup, I needed a negative offset with BIOS 4205, but with the latest BIOS you might need a positive one—though that's still possible with your model. Once you find the right Vcore, expect voltage fluctuations between 1.44V and 1.45V, with single-core boosts possibly matching those levels. Ryzen handles spikes up to 1.55V and can sustain around 1.5V, but optimizing it should keep voltages within a safe range—no more than 1.46V in boost mode, with occasional brief spikes of 1.5V.

J
Jayhawk_Down
Senior Member
350
03-07-2020, 06:51 AM
#25
I just wanted to share a quick update—I noticed today that Asus updated their BIOS to version 4804 on their website.
ROG STRIX X470-F GAMING | Motherboards | ROG Global
This appears to be the top AMD Ryzen ATX motherboard with Aura Sync, SupremeFX, ROG Audio, Dual M.2, Intel LAN, VR support, M.2 heatsink, and USB 3.1 Gen 2.
www.asus.com
After flashing the new BIOS, I spent the day testing it. It seems to offer slightly improved memory stability, allowing me to fine-tune my timings a bit more. On the CPU side, I now only need a positive offset of 0.03125V to achieve full core boost at 4.35Ghz (with PBO). Each processor and BIOS version behaves differently regarding voltage needs, but every update brings changes. With this latest BIOS, I actually need less voltage for full stability. The process took several hours to get the right settings, but it’s definitely worth the effort.
J
Jayhawk_Down
03-07-2020, 06:51 AM #25

I just wanted to share a quick update—I noticed today that Asus updated their BIOS to version 4804 on their website.
ROG STRIX X470-F GAMING | Motherboards | ROG Global
This appears to be the top AMD Ryzen ATX motherboard with Aura Sync, SupremeFX, ROG Audio, Dual M.2, Intel LAN, VR support, M.2 heatsink, and USB 3.1 Gen 2.
www.asus.com
After flashing the new BIOS, I spent the day testing it. It seems to offer slightly improved memory stability, allowing me to fine-tune my timings a bit more. On the CPU side, I now only need a positive offset of 0.03125V to achieve full core boost at 4.35Ghz (with PBO). Each processor and BIOS version behaves differently regarding voltage needs, but every update brings changes. With this latest BIOS, I actually need less voltage for full stability. The process took several hours to get the right settings, but it’s definitely worth the effort.

Pages (3): Previous 1 2 3