Any idea on a vague timeline for AGESA 1007 to hit public beta for the K7?
I'm assuming here, but I would imagine we've seen a delay the past couple of weeks on the BIOS update side due to your team focusing mostly on a BIOS for 1007. Given how extensive the changes in 1007 apparently are, it would seem like a smart decision, but I think it would help a lot of us to know that this is the case. IDK if you can share such information, but a very rough idea of when we might expect a 1007 BIOS would still be nice.
Mid/End of November for new AGESA. From what I hear they are going to call it 1070 not 1007 - Don't ask me why.
Right now I'm Offline I have made 7 posts
I joined September 2017
I've received 1 like
My gender is Male
I currently have no messenger contact information filled out.
I currently have no social network information filled out.
Built a PC for customer with AX370 Gaming K5 bios. Shocked at how bad this bios is. It's something you'd expect to see on a $50 motherboard. It's like a bios for children no more than 6 years of age. No LLC, no fixed voltages and when I change the CPU multiplier in the bios it's basically ignored and the motherboard keeps the Ryzen 1700 at stock speed. If I go into windows and use the Giga tuner software then the overclock works... Obviously an overclock that can only be set from windows by loading giga software and then clicking on the profile for the overclock is useless.
Built a PC for customer with AX370 Gaming K5 bios. Shocked at how bad this bios is. It's something you'd expect to see on a $50 motherboard. It's like a bios for children no more than 6 years of age. No LLC, no fixed voltages and when I change the CPU multiplier in the bios it's basically ignored and the motherboard keeps the Ryzen 1700 at stock speed. If I go into windows and use the Giga tuner software then the overclock works... Obviously an overclock that can only be set from windows by loading giga software and then clicking on the profile for the overclock is useless.
The K5 doesn't have digital voltage control, thus you can only do offsets. You should however be able to change multiplier no problem. I assume you used the latest BETA BIOS? Did you try a previous BIOS?
Built a PC for customer with AX370 Gaming K5 bios. Shocked at how bad this bios is. It's something you'd expect to see on a $50 motherboard. It's like a bios for children no more than 6 years of age. No LLC, no fixed voltages and when I change the CPU multiplier in the bios it's basically ignored and the motherboard keeps the Ryzen 1700 at stock speed. If I go into windows and use the Giga tuner software then the overclock works... Obviously an overclock that can only be set from windows by loading giga software and then clicking on the profile for the overclock is useless.
That's a bug specific to your board or bios, multiplier works just fine on my K5 with latest stable bios, and I think it worked fine with earlier bios as well. May have to reflash your bios or reseat your cpu (I've heard you need to press down on the heatspreader when you lock the cpu in for a good seat, after letting it drop in naturally of course. A bad seat has caused a few problems reported on these forums).
5. Constant disconnecting with external USB 3.0 HDDs when connected to the front USB ports in Windows 10.**
** I tried 3 different external HDDs all of which have the same problem. They work perfectly on any other system I use them with. The problem is most easily replicated when transferring a large amount of data (over 100GB) to the external drive or when playing audio / video content from the external drive, it will just randomly disconnect. This does not happen on any of the 3 external drives I tried on any other system I use.
Is it only on the front panel? If so it maybe a voltage issue. You can use the DAC-UP software to raise the voltage and see if solves the issue.
I heard somewhere that enabling(?) the USB MSI (Message Signal Interrupt) option can fix this. This is one of the many advanced options that Gigabyte has left hidden.
My Signature... AMD Ryzen 7 1700 @ 3.8GHz, 1.25625v AX370 Gaming 5 @ UEFI version F22b modded by ket Corsair Vengeance LPX CMK32GX4M2B3200C16R Ver. 5.39 (2x16GB Hynix DR) @ 2933, 14-16-16-33-53-1T Sapphire RX560 Pulse 4GB
We implemented P-state overclocking on our K7. So far the feedback I have heard in this thread is mostly positive with some quirks... We are still evaluating if we will trickle it down to other boards. Many other manufacturers implemented it, then pulled it. AFAIK very few boards have this option - mostly because its not a huge usage case and because it can cause boot issues.
Hey Matt, can we get P-State bug where setting VID value that results voltage > 1.35V make that P-State invalid (CPU just switches to next p-state) fixed? Right now ppl have to leave VID at default and use offset voltage as work around instead.
We implemented P-state overclocking on our K7. So far the feedback I have heard in this thread is mostly positive with some quirks... We are still evaluating if we will trickle it down to other boards. Many other manufacturers implemented it, then pulled it. AFAIK very few boards have this option - mostly because its not a huge usage case and because it can cause boot issues.
I hope Gigabyte won't drop P-state OC. I really want to implement but waiting on a New Bios that is just around the corner...
1. There are no P-State options for P-State overclocking. Every other manufacturer has this on their lowest end boards, the Gaming 5 not having it at all is a complete joke.
We implemented P-state overclocking on our K7. So far the feedback I have heard in this thread is mostly positive with some quirks... We are still evaluating if we will trickle it down to other boards. Many other manufacturers implemented it, then pulled it. AFAIK very few boards have this option - mostly because its not a huge usage case and because it can cause boot issues.
P-State overclocking is quite essential to have cool running but performing systems.
The 1700x on an AB350 Gaming 3 seems to expose P0 / P1 / P2 - and I use the ZenStates-Linux scripts to change these at boot time.
I set the following: ### The 1700x only seems to support P0/1/2 ## Defaults are: # P0 - Enabled - FID = 88 - DID = 8 - VID = 20 - Ratio = 34.00 - vCore = 1.35000 # P1 - Enabled - FID = 78 - DID = 8 - VID = 2C - Ratio = 30.00 - vCore = 1.27500 # P2 - Enabled - FID = 84 - DID = C - VID = 68 - Ratio = 22.00 - vCore = 0.90000
The CPU then follows these P-States without issue.
I did note that it I absolutely torture the system, to get all 8 cores / 16 threads maxed out at 3.9Ghz during the stress tests, I need to bump the P0 voltage to 1.375v - however under normal usage of the system, this isn't needed.
If I could do this straight in the BIOS instead of with third party tools after the fact (and no, Ryzen Master doesn't do this), then I could use these settings in both Linux and Windows.
Without this support, I'm stuck at setting the voltage at 1.35v all the time in Windows.
For what its worth, it looks like the 1700x *may* support P0-P7 - and I can enable them using the same utility - but when I reboot, the BIOS complains about incompatible settings in the CPU and won't work until a hard reset... See the following:
# ./zenstates.py -l P0 - Enabled - FID = 9C - DID = 8 - VID = 20 - Ratio = 39.00 - vCore = 1.35000 P1 - Enabled - FID = 94 - DID = 8 - VID = 28 - Ratio = 37.00 - vCore = 1.30000 P2 - Enabled - FID = 66 - DID = C - VID = 78 - Ratio = 17.00 - vCore = 0.80000 P3 - Disabled P4 - Disabled P5 - Disabled P6 - Disabled P7 - Disabled C6 State - Package - Enabled C6 State - Core - Enabled
EDIT: For calculating the values, I use this spreadsheet.RyZenPstateCalc.xlsx (296.82 KB)
My Signature... AMD Ryzen 1700x Gigabyte AB350 Gaming 3 - BIOS F9d Gigabyte GeForce® GTX 1060 G1 Gaming 6G G.Skill Ripjaws V F4-2400C15D-16GVR 16GB (2x8GB) 2400MHz DDR4 Samsung 960 EVO 500GB PCIe SSD Corsair H110i CPU Cooler Thermaltake ToughPower Gold 750W Power Supply Phanteks Eclipse P400S Satin Black Tempered Glass ATX Case
We implemented P-state overclocking on our K7. So far the feedback I have heard in this thread is mostly positive with some quirks... We are still evaluating if we will trickle it down to other boards. Many other manufacturers implemented it, then pulled it. AFAIK very few boards have this option - mostly because its not a huge usage case and because it can cause boot issues.
Hey Matt, can we get P-State bug where setting VID value that results voltage > 1.35V make that P-State invalid (CPU just switches to next p-state) fixed? Right now ppl have to leave VID at default and use offset voltage as work around instead.
Yes please fix this. Getting P0 being skipped to P1 after a certain point. Mine is anything over 3.65 overclock using P-State. Probably same issue, but mine is happening lower than 1.35 VID.
Also this beta screws my grub2 linux EFI partition (minor issue, easy fix). Further probably related to same EFI issue, PS2 keyboard doesnt register early enough in boot sequence so I cannot pres F12 to choose boot option. EFI USB is also making new boot EFI entry after restart, resulted in 28 EFI partition of same partition entry.
Don`t talk about this ket guy he is some retard.. How i know easy go find him and his retarded review of x370 Aorus 5 on extreme systems, and you will know for what im talking about, he just don`t have a ****king clue what he is talking about.. And other thing he is NOT modding these bioses, he is just unlock hidden (i assume from Gigabyte for a reason) options and make them visible, nothing more...
Better not to trust such a uneducated little kid....
My Signature... AMD Zen 1700+ - 3.9GHz Aorus x370 K7 - Thermal Armor G. Skill Ripjaws V - 16GB 3200MHz HyniX M-Die Future Radeon RX Vega or GTX 1080TI EVGA SuperNOVA 850W Platinum Fractal Design Define R2 XL Black Pearl Noctua D15 Premium Cooler IBM Model M - 1391403
Right now I'm Offline I have made 123 posts
I joined June 2017
I've received 6 likes
My gender is Male
I currently have no messenger contact information filled out.
I currently have no social network information filled out.
P-States should be available on all board, simply because it have other bug (in AGESA I think) that with some really good binned CPU (with lower VID by default) will get stuck in P2 if you offset up the voltage with lead to 1.55Ghz on non-X and 2.2Ghz on X CPU. The only way to bypass that bug is to use p-states OC.
For me it's not really a deal breaker as I don't use Windows and I added some ASM instruction in my boot loader to overwrite the CPU p-states (I did use zenstates to test first then modified grub and xen to get it patched before the OS boot to get accurate reading)
But for others who use Windows or don't want to patch the code, they should be able to overclock properly, especially with a good binned CPU, where that bug hit the most....
4Ghz @ 1.33V, stable (AMD send good CPU when you RMA for that segmentation fault bug)
We appreciate the feedback. I can assure you I have read this entire thread as well as its predecessor more than a few times. I collect your guys input/feedback and loop it back to our R&D/BIOS teams. We have put out a pretty hefty amount of BETA bios since launch and will continue to improve/refine the BIOS as we fix bugs and new AGESA is released by AMD. As I have stated in the past AM4 has turned out to be alot of 2 steps forward one step back. Lets take a moment to realize some of the milestones in memory alone we have seen:
Originally memory was locked down with very few multipliers. (IIRC 3200 was the max?) The only way to increase this was BLCK... After alot of pushing we now have multipliers to match today's DDR4!
Along with few options for memory multipliers we started struggling to get any memory working! Now the struggle is getting 3200 to be plug and play (We will get there!)
Long story short just remember this is new (and exciting!) territory in many ways.... It's going to have quirks and we will continue to support you guys in ironing them out.
1. There are no P-State options for P-State overclocking. Every other manufacturer has this on their lowest end boards, the Gaming 5 not having it at all is a complete joke.
We implemented P-state overclocking on our K7. So far the feedback I have heard in this thread is mostly positive with some quirks... We are still evaluating if we will trickle it down to other boards. Many other manufacturers implemented it, then pulled it. AFAIK very few boards have this option - mostly because its not a huge usage case and because it can cause boot issues.
2. Voltage readings on the board are extremely erratic even when using various levels of LLC (I found the most stable setting to be "High"). Eg; CPU voltage bouncing between 1.33v and 1.39v under load. Recalibration looks necessary here at least because that voltage stability is just awful.
I can pass this feedback along, however my understanding is LLC in relation to Vcore is not static. I can double check but I believe extreme is more of a flatline. The trade off here is the voltage will be higher at all times, rather than dip and rise. You can try lowering Vcore and running extreme if you want less fluctuation.
3. 2T command rate does not work when setting it in the memory timing options or when Gear Down mode is disabled.*
Last I talked to AMD 2T is actually more of a placebo. Internally the memory runs at 1T regardless. So anytime you see "2T" its still in "1T" with a work around (gear down) in reporting.
* I Tried the dividers and command rate with the XMP profile enabled and disabled as well as switching the DIMM banks the memory is in. In all cases the result is the same; no POST. The memory I have uses Samsung B-Die ICs and the 2133, 2800, 2933, 3066, 3200, 3333, etc dividers work fine.
5. Constant disconnecting with external USB 3.0 HDDs when connected to the front USB ports in Windows 10.**
** I tried 3 different external HDDs all of which have the same problem. They work perfectly on any other system I use them with. The problem is most easily replicated when transferring a large amount of data (over 100GB) to the external drive or when playing audio / video content from the external drive, it will just randomly disconnect. This does not happen on any of the 3 external drives I tried on any other system I use.
Is it only on the front panel? If so it maybe a voltage issue. You can use the DAC-UP software to raise the voltage and see if solves the issue.
6. When selecting the Orange colour profile in RGB Fusion the LED colour is actually yellow. It should be Red:255 Green:20 Blue:0. This should surely be an easy fix as someone just needs to alter a couple characters of code correcting the preset values.
7. Not able to save UEFI profiles to removable media such as USB sticks. The wrong option is displayed in the "Load Profiles" menu. Its the one to load from HDD/FDD/USB that appears.
8. There are no Southbridge options, at all. The Gaming 5 firmware is also missing quite a few other options some users at least I'm sure would find useful upon me doing some googling and comparing it to the likes of Asus and Asrock.
Which options are you referring? "Southbridge" does not exist on AM4, all functions are routed through the CPU or chipset.
I'll start top to bottom for completeness but before getting in to that I have also recently discovered that the Gaming 5 when overclocking with the multiplier never throttles the CPU speed down. This is (again) a bug I have only encountered with the Gaming 5 and none of the other AM4 boards I have tested with.
1. I have now used the Asrock Fatal1ty Gaming K4, Asrock Fatal1ty Gaming X, Asus ROG Crosshair VI Hero, and Asrock X370 Taichi in testing my Samsung B-Die and at time of testing a couple days ago all of those boards have P-State overclocking options in the UEFI they were using which admittedly I did not check as my only concern was with memory testing but the boards were from new shipments so bound to have fairly recent firmware on at least. I also never experienced any issues with P-State overclocking in the quick tests I did with those boards. As far as I am concerned as other manufacturers are already including P-State overclocking in boards that are less expensive than the Aorus Gaming 5 (Gaming K4 and Gaming X, for example) there is no "evaluating" to be done here it is a foregone conclusion to include the capability and maintain a equally competitive playing field. Not to do so for the likes of the Gaming 5, K7 and K5 at least would be bad form.
2. All the boards I used testing my Samsung B-Die AFAIK measure and show voltages in realtime the Aorus Gaming 5 just has by far the most unstable voltage regulation I've seen compared to all the other AM4 boards I tested my memory on. The other boards had minor fluctuation of about 0.013v this Gaming 5 fluctuates as much as 0.062v with average fluctuation of 0.050v which is just bad, really bad. Theres no other way to say that.
3. 2T command rate, or Gear Down as you may prefer to call it, is not placebo with AM4. I can disable Gear Down on any of the other boards I tested my Samsung B-Die with and at the very least have increased stability which I have verified by using HCI Memtest. The difference is errors occurring instantly or after a few minutes (depending on tested frequency) with Gear Down enabled (1T), to increasing to several hours of stability with Gear Down disabled (2T). I further verified this was not a fluke by also running an AIDA bandwidth test, when disabling Gear Down bandwidth did decrease by 2-3GB which is exactly what you would expect when switching from a 1T command rate to 2T.
4. What I said with point 1 covers this really, all the other boards work fine with the 2400 and 2666 memory dividers its just the Aorus Gaming 5 that throws a fit and refuses to POST but 2133, 2800, 2933, 3200, etc dividers all work. Very odd.
5. Yes its just the front panel. I have already tried increasing USB voltage along with a multitude of other things such as making sure I have the latest Asmedia drivers (v1.16.47.2) all the latest Windows updates, latest drivers from AMD etc, and absolutely nothing has worked.
6 / 7. No other info I can give you here its all exactly as it says on the tin, the Orange LED profile is yellow where the preset is wrong and when trying to save a UEFI profile to USB/HDD etc the wrong option is displayed. Below is a screenshot of that for some added clarity.
8. By "southbridge" I should refer to it as chipset, old habits I guess. Some of the options I'm talking about are as follows; chipset power saving features, spread spectrum options, USB phy power down, USB MSI, and theres one or two other options I've seen people asking about and being left rather perplexed by Gigabyte not including them particularly in the Linux community where security is taken extremely seriously and the lack of these options does put a black mark on Gigabyte for not including them. One such option that has just sprung back to mind is ATA security.
Nothing else comes to mind right now except for the generally clunky and poor layout of the UEFI. If I encounter other problems I'll put them in this thread.
Right now I'm Offline I have made 40 posts
I joined June 2017
I've received 1 like
My gender is Male
I currently have no messenger contact information filled out.
I currently have no social network information filled out.
After my last post while ago I was still getting BSODs all the time with 2x16GB Corsair Vengeance LPX (CMK16GX4M1B3000C15, v5.39). Nothing helped, BSOD mostly while idle on desktop. So I raised DRAM voltage even further - to 1.39 or 1.4V. Still reading in every tool including SIV is higher, 1.416V at the moment. Running whole day without BSOD at 2933.
Please check why voltages for RAM in reality are higher than what is set in UEFI on Gaming 5, F9a bios.
I think that wasn't the case in earlier versions. Now when the thing finally works I am afraid to try F9c or change anything:)
What are safe voltages for 24/7 on DDR4 because I find different information. 1.45V OK? Asking just in case I have to add 0.1 or 0.2V more to 1.42.
Is it, with raising DRAMV (more than 1.4V), necessary to bump up SOC voltage further than current 1.1250V or DRAM Termination to 0.700-0.805 (half of voltage)?
Does anybody else have this problem, with higher DRAMV on Gaming 5 with F9a BIOS?
Right now I'm Offline I have made 56 posts
I joined April 2017
I've received 6 likes
My gender is Male
I currently have no messenger contact information filled out.
I currently have no social network information filled out.
6 / 7. No other info I can give you here its all exactly as it says on the tin, the Orange LED profile is yellow where the preset is wrong and when trying to save a UEFI profile to USB/HDD etc the wrong option is displayed. Below is a screenshot of that for some added clarity.
Well orange is orange by me And the select option is not wrong but just says as it is ,select a device to save to
helpful55 wrote: Dec 14, 2019 16:14:20 GMT -8 have my pc broke and no help for weeks. (And it would be nice to be able to put a full comment/thought here.)
helpful55 wrote: Dec 14, 2019 16:12:35 GMT -8 Unbelievable that such a large company can’t run the essentials for we paying customers. What are the dates that support will be unavailable?
helpful55 wrote: Dec 14, 2019 16:12:03 GMT -8 How do I contact giga? It says the contact is 'under maintenance', and support the same. Facebook page is gone. I need to work on my pc and I am trying to plan because I know they take off an awful lot of time for what they want for holidays and I can't ha
1spidee wrote: Dec 12, 2019 16:28:26 GMT -8 I have an Aorus b450 M amd stock fan. Any suggestions for a better air cooler cpu 2700
bigdave wrote: Dec 11, 2019 19:51:11 GMT -8 davegps, Please start a thread on the forum.
davegps wrote: Dec 11, 2019 13:47:14 GMT -8 Anyone else have an issue with the poor Aorus support ? I two support requests for a AORUS-Gen4-AIC-SSD-8TB SSD, one for it only showing 2TB, the software for configuring it can't find the card, however the SSD Tool Box software can....I can't even regist
burningsparks wrote: Dec 10, 2019 18:29:58 GMT -8 I have built a PC and whenever i connect it to the monitor it says no signal, i have tried two different monitors but no BIOS boot screen. Connected to monitor using motherboards HDMI port.
The CPU/DRAM/VGA/BOOT LED indicators seems to work fine but no s
jiml wrote: Dec 8, 2019 14:53:07 GMT -8 Hi...Need help with my new Build - Aorus x570 Ultra MOBO, Patriot Viper 3200 2x16 GB DDR4 RAM (On approved List) Noctua CPU Fan, ...won't post...red light on MOBO...Thanks!
amraxjane wrote: Dec 8, 2019 9:30:21 GMT -8 How to solve crack/noise issue of speakers when the pc is power on , mobo GA H270 Gaming 3
memeisms wrote: Dec 6, 2019 14:31:56 GMT -8 Do the UD line of Z390 mobos not support Wake-On-Lan? I cannot for the life of me find Wake-On-Lan settings in my Z390 UD rev1.0 BIOS. Someone please help lol. EDIT: On latest BIOS version, F9.*
lebobbi wrote: Dec 3, 2019 17:36:36 GMT -8 RGB Fusion does not start anymore not even after uninstalling, restarting resintalling.
czagarskas wrote: Dec 2, 2019 21:17:34 GMT -8 Using f10 bios, newest one, and the Aorus X570 PCH fan still has no control. I have a custom cooling system in here, I need the fan to be OFF when I am recording audio. Anyone figured this out yet?
mckootenay wrote: Nov 29, 2019 7:49:01 GMT -8 No error code on blue screen in event viewer and still happens after reinstall win10 or replacing M.2 drive
mckootenay wrote: Nov 29, 2019 7:48:19 GMT -8 Wierd issue with B450 using M.2 SSD only. Boots into win10 after a few minutes it blue screens and reboots. then it is fine for the rest of the time. sometimes locks at Aorus screen or win logo as well. after the reboot it is all ok though
jp96785 wrote: Nov 28, 2019 20:51:39 GMT -8 it isnt too bad
(AMD send good CPU when you RMA for that segmentation fault bug)
