Yeah, and is there a possibilty that you could share the 0505 BIOS file?
AMD Ryzen 7 1800X CPU Review
- Thread starter tinytomlogan
- Start date
Yeah, and is there a possibilty that you could share the 0505 BIOS file?
OMG! I never thought that you would share it, im so gratefull, thanks!!! I'll share the info if my "old" RAM works better with it and if there are any changes a.s.o.
R7 1800X temperature & SMT efficiencies ?
My correspondence here is a continuation of my comments of 05-03-2017 3:21 pm. I have had some time to reflect and think about your test data and your video testing of the Ryzen 1800X with the ASUS Prime X370 pro motherboard.
Observation in regards to the temperature noted with the UEFI ASUS bios utility and the AMD Ryzen master control panel and providing CPU loading with OCCT 4.5.0 software. From your video the temperature were noted about 80 degree C with an overclock of about 4 GHz for the CPU. The test circuit had a liquid cooling CPU interface with radiator bank and two 140 mm fans. The desktop also had an external rear exhaust port fan (140 mm?) and appears two internal fans at the front of the computer case (120 mm?). My question is related to the CPU temperature.
From experience with an ASUS Sabretooth 990FX motherboard the software for the bios and the AI suite II can display the CPU temperature. I have found thru investigation the temperature is not actually the internal temperature of the CPU but rather the CPU socket temperature. I have confirmed the correct CPU temperature with HWINF64 software and software supplied with my liquid radiator (NXZT Kraken X61). I also checked with a thermal couple on the CPU socket to confirm the HWINF64 software was indeed correct. I am wondering if the new ASUS Prime X370 motherboard and AMD's master control panel software also are measuring the CPU socket temperature rather than the actual CPU internal temperature. As we are testing a new platform and software this can easily be misconstrued and review of the vendors motherboard guide are not clear. Could you please review this matter and clarify. Caught my attention when Mr. Logan placed his hand over the radiator and touched the surface of the radiator and very little heat was noted by him.
Furthermore if the temperatures are quite high for the socket I am wondering because of the test circuit set up that the fan air flow is not balanced to circulate the air around the CPU socket. Could you please check to see what sort of air is flowing near the surface of the pump to socket interface and also on the back side of the motherboard around the CPU interface. From experience I have seen air circulation issues with radiator fans that create a negative air pressure within the desktop (more air exhausted then entering) and poor air circulation near the motherboard to CPU socket interface, and the VRM, resulting in higher socket temperatures. As the AM4 is a new platform and the pump to socket interface is quite large, might be some restriction of air circulation in the vicinity of the CPU socket because of fan air flow balance is improperly set, or insufficient air movement along the surface of the motherboard. Again the AM4 is a new platform including the software requires a bit of a learning curve and the efforts by Mr. Logan during testing are appreciated. Please review.
My next observation is in regards to the synthetic and real benchmark testing of the Ryzen 1800X CPU. The results are very interesting and the efforts by OC3D team are noted. What caught my attention is that Intel is still a step ahead at single thread processing over AMD. AMD has made great stride in single thread processing and results are very good. What is really interesting is AMD's multi-threading processing is beating Intel in the synthetic and real benchmark testing even with Intel's HT faster at thread processing than SMT by AMD. So the question is WHY? Well the data provided by Mr. Logan has the answers but can be a little bit hard to see the Tree's thru the Forrest. The data is illustrating that SMT, a new and completely different technology than Intel's HT is more efficient at processing the multi thread tasking. The efficiency can be easily demonstrated by repeating some of the synthetic and real benchmark testing but this time setting the R7 1800X, i7 7700K, i7 6900K, i7 6950X all at the same CPU frequency say 3.5, 3.8 or 4.0 GHz and compare the benchmarks using a percent scaling. This method of testing eliminates frequency scaling and provides efficiency comparisons of the CPU's. The data anticipated would illustrate that the AMD SMT technology is more efficient than Intel's HT somewhere in the range of about 5 to 25 percent depending on the Intel's CPU and likely at par or slightly better that Intel’s 10 core CPU i7 6950X during the bench marking!
Last the engineers at AMD are congratulated for efforts in developing the SMT technology and the new Ryzen CPU's. AMD efforts have opened up a competitive market and moved the CPU technology forward for everyone. Thank you OC3D for your test review and your staff's input.
Best regards
Terry
My correspondence here is a continuation of my comments of 05-03-2017 3:21 pm. I have had some time to reflect and think about your test data and your video testing of the Ryzen 1800X with the ASUS Prime X370 pro motherboard.
Observation in regards to the temperature noted with the UEFI ASUS bios utility and the AMD Ryzen master control panel and providing CPU loading with OCCT 4.5.0 software. From your video the temperature were noted about 80 degree C with an overclock of about 4 GHz for the CPU. The test circuit had a liquid cooling CPU interface with radiator bank and two 140 mm fans. The desktop also had an external rear exhaust port fan (140 mm?) and appears two internal fans at the front of the computer case (120 mm?). My question is related to the CPU temperature.
From experience with an ASUS Sabretooth 990FX motherboard the software for the bios and the AI suite II can display the CPU temperature. I have found thru investigation the temperature is not actually the internal temperature of the CPU but rather the CPU socket temperature. I have confirmed the correct CPU temperature with HWINF64 software and software supplied with my liquid radiator (NXZT Kraken X61). I also checked with a thermal couple on the CPU socket to confirm the HWINF64 software was indeed correct. I am wondering if the new ASUS Prime X370 motherboard and AMD's master control panel software also are measuring the CPU socket temperature rather than the actual CPU internal temperature. As we are testing a new platform and software this can easily be misconstrued and review of the vendors motherboard guide are not clear. Could you please review this matter and clarify. Caught my attention when Mr. Logan placed his hand over the radiator and touched the surface of the radiator and very little heat was noted by him.
Furthermore if the temperatures are quite high for the socket I am wondering because of the test circuit set up that the fan air flow is not balanced to circulate the air around the CPU socket. Could you please check to see what sort of air is flowing near the surface of the pump to socket interface and also on the back side of the motherboard around the CPU interface. From experience I have seen air circulation issues with radiator fans that create a negative air pressure within the desktop (more air exhausted then entering) and poor air circulation near the motherboard to CPU socket interface, and the VRM, resulting in higher socket temperatures. As the AM4 is a new platform and the pump to socket interface is quite large, might be some restriction of air circulation in the vicinity of the CPU socket because of fan air flow balance is improperly set, or insufficient air movement along the surface of the motherboard. Again the AM4 is a new platform including the software requires a bit of a learning curve and the efforts by Mr. Logan during testing are appreciated. Please review.
My next observation is in regards to the synthetic and real benchmark testing of the Ryzen 1800X CPU. The results are very interesting and the efforts by OC3D team are noted. What caught my attention is that Intel is still a step ahead at single thread processing over AMD. AMD has made great stride in single thread processing and results are very good. What is really interesting is AMD's multi-threading processing is beating Intel in the synthetic and real benchmark testing even with Intel's HT faster at thread processing than SMT by AMD. So the question is WHY? Well the data provided by Mr. Logan has the answers but can be a little bit hard to see the Tree's thru the Forrest. The data is illustrating that SMT, a new and completely different technology than Intel's HT is more efficient at processing the multi thread tasking. The efficiency can be easily demonstrated by repeating some of the synthetic and real benchmark testing but this time setting the R7 1800X, i7 7700K, i7 6900K, i7 6950X all at the same CPU frequency say 3.5, 3.8 or 4.0 GHz and compare the benchmarks using a percent scaling. This method of testing eliminates frequency scaling and provides efficiency comparisons of the CPU's. The data anticipated would illustrate that the AMD SMT technology is more efficient than Intel's HT somewhere in the range of about 5 to 25 percent depending on the Intel's CPU and likely at par or slightly better that Intel’s 10 core CPU i7 6950X during the bench marking!
Last the engineers at AMD are congratulated for efforts in developing the SMT technology and the new Ryzen CPU's. AMD efforts have opened up a competitive market and moved the CPU technology forward for everyone. Thank you OC3D for your test review and your staff's input.
Best regards
Terry
So, after a night of testing with 0505 vs 0504 BIOS *jawn* the results:
RAM:
"old"
GeiL Evo Potenza Quad-Kit 2400 MHz 15-15-15-35-2T 32 GB (4x 8) (Dual-Rank)
"new"
G.Skill Aegis Dual-Kit 2400 MHz 15-15-15-35-1T 16 GB (2x 8) (Single-Rank)
Standard voltages are all the same, temprature reading is also the same.
So far so good but will it take my "old" quad-kit?
No, both refuse to run stable, with the 0504 BIOS the startup is okay, then in windows the system is instable and gets random bluescreens.
Tried 1866, 2133 and 2400 MHz, losen the timings, nothing works.
With the 0505 no boot not even with only 2 DIMMs installed.
Kinda depressing and it was worth a try but meh, hope dies last ^^
So next was testing if my "new" RAM gets above 2400 MHz. It was rockstable at 2400 MHz D.O.C.P. 15-15-15-35-1T so I've upped it to 2666, upped the SOC voltage to 1.0 (default/auto is 0.9) and the DRAM voltage to 1.25... It booted, it worked in windows but under stress it crashed with BSOD, so I've upped the timings to 16-16-15-38-1T and yeah 2 hours RealBench stresstest with 16 GB setting, 1h Prime95 with max RAM usage - worked like a charm.
With the 0504 BIOS 2666 MHz wasn't working under the same circumstances.
Conclusion:
So maybe Asus tweaked a 0504 BIOS for overclocking RAM?
Because when you flash the 0504, or the 0505 BIOS both have the same creation date: 02/27/2017
And the Prime x370 Pro has maybe some problems, in it's actual BIOS state, with dual-rank DIMMs as what I've read I'm not the only one who ran into this kind of issue with the mobo.
Otherwise, when you got compatible RAM/DIMMs (QVL - paid list *cough*) it seems to be rock solid: 1700 (non X) @ 3,8 GHz with 1.34 voltage (LLC 1) and RAM at the speed above.
So I hope this could help someone who had/has the same problem.
RAM:
"old"
GeiL Evo Potenza Quad-Kit 2400 MHz 15-15-15-35-2T 32 GB (4x 8) (Dual-Rank)
"new"
G.Skill Aegis Dual-Kit 2400 MHz 15-15-15-35-1T 16 GB (2x 8) (Single-Rank)
Standard voltages are all the same, temprature reading is also the same.
So far so good but will it take my "old" quad-kit?
No, both refuse to run stable, with the 0504 BIOS the startup is okay, then in windows the system is instable and gets random bluescreens.
Tried 1866, 2133 and 2400 MHz, losen the timings, nothing works.
With the 0505 no boot not even with only 2 DIMMs installed.
Kinda depressing and it was worth a try but meh, hope dies last ^^
So next was testing if my "new" RAM gets above 2400 MHz. It was rockstable at 2400 MHz D.O.C.P. 15-15-15-35-1T so I've upped it to 2666, upped the SOC voltage to 1.0 (default/auto is 0.9) and the DRAM voltage to 1.25... It booted, it worked in windows but under stress it crashed with BSOD, so I've upped the timings to 16-16-15-38-1T and yeah 2 hours RealBench stresstest with 16 GB setting, 1h Prime95 with max RAM usage - worked like a charm.
With the 0504 BIOS 2666 MHz wasn't working under the same circumstances.
Conclusion:
So maybe Asus tweaked a 0504 BIOS for overclocking RAM?
Because when you flash the 0504, or the 0505 BIOS both have the same creation date: 02/27/2017
And the Prime x370 Pro has maybe some problems, in it's actual BIOS state, with dual-rank DIMMs as what I've read I'm not the only one who ran into this kind of issue with the mobo.
Otherwise, when you got compatible RAM/DIMMs (QVL - paid list *cough*) it seems to be rock solid: 1700 (non X) @ 3,8 GHz with 1.34 voltage (LLC 1) and RAM at the speed above.
So I hope this could help someone who had/has the same problem.
Avet
Active member
In Air 540 chase with 2x140 fans on AIO running at 100%, and 3x120mm chase fans no matter what direction you put them on you won't get stagnant air around CPU socket to affect temperatures. There is nothing in the main chamber to disturb the air current. And the rear fan is blowing in over the CPU.
Understand that this CPU is like Bugati Veyron, basically W16 engine. Regular cars need 1 radiator to cool them, Veyron needs 10. More cylinders (cores)=more heat. Simple as that. This is a proper beast under the hood so please stop commenting temperatures. They are fine. It is meant, and built to run that hot. It won't bother him.
Last edited:
BigDaddyKong
Active member
https://community.amd.com/community/gaming/blog/2017/03/13/amd-ryzen-community-update?sf62107357=1
20C offset, not bad. Temps are much better than people thought.
Last edited:
NeverBackDown
AMD Enthusiast
Its just upto monitoring software to catch up now. Otherwise you will have really high fan speeds for no reason.
BigDaddyKong
Active member
I am more worried about
A legitimate reason to go back to Windows 7?
A legitimate reason to go back to Windows 7?
Just something on this, does anyone else find it odd as hell that they are adding 20 degrees for a fan setting. Someone explain that to me. As it stands we cant overclock more because it thinks its 20 degrees hotter than it is.
Now colour me silly but isnt this the reason we have fan curves and custom water cooling, what is the point of doing any of that if its throttling way before it even gets close to its limit.
And no wonder tom was saying the rads were cold when it was saying it was hot, these things have so much overhead that is simply not being touched because the systems are thinking there frying when they arnt.
BigDaddyKong
Active member
TTL, please fire up Win 7 and Win 10 and give us a comparison.
Now we know we have room to really pound on these babies with some volts and proper cooling.
Now we know we have room to really pound on these babies with some volts and proper cooling.
BigDaddyKong
Active member
Most temp monitoring programs have a place for off set. You only need to set it to -20C and it will read correctly. In a month or two, programs will update and give you the correct temps automatically.
Well im leaning more towards the bios, doesnt make a dif if the software is saying the correct temp if the bios is still pulling back and even shutting down because it thinks the thing is 20 degrees hotter then it is.
Makes watercooling kind of mute.
Following OC3D for some time. Now I had to register.
How did you get this system to run stable?
I'm building a system https://pcpartpicker.com/list/J4CFqk using same MB. But even after updating all firmware (really helped the 960 EVO) and drivers I still frequently get BSODs like IRQL_NOT_LESS_OR_EQUAL or KMODE_EXCEPTION_NOT_HANDLED.
I used HX424C15FBK2/16 dimms found on mobos QVL.
Yes yes ok, GPU is old.
Have been building mine and others pcs since 80486's this is the first time firmware/driver updates or bios/autoexec/config.sys tinkering didn't solve the problem.
How did you get this system to run stable?
I'm building a system https://pcpartpicker.com/list/J4CFqk using same MB. But even after updating all firmware (really helped the 960 EVO) and drivers I still frequently get BSODs like IRQL_NOT_LESS_OR_EQUAL or KMODE_EXCEPTION_NOT_HANDLED.
I used HX424C15FBK2/16 dimms found on mobos QVL.
Yes yes ok, GPU is old.
Have been building mine and others pcs since 80486's this is the first time firmware/driver updates or bios/autoexec/config.sys tinkering didn't solve the problem.
Last edited:
Does anyone have any info on the 0505 BIOS? Doubtful but thought to ask anyway.
I have the 0504 BIOS installed, ram running at 2933Mhz and a modest CPU overclock to 3.87Ghz. I can run my ram at 3200 but i get system instability when attempting to overclock the cpu past its base settings of 3.6Ghz.
If theres no details on the BIOS, i will leave it as it is and wait for official BIOS updates.
Cheers
I have the 0504 BIOS installed, ram running at 2933Mhz and a modest CPU overclock to 3.87Ghz. I can run my ram at 3200 but i get system instability when attempting to overclock the cpu past its base settings of 3.6Ghz.
If theres no details on the BIOS, i will leave it as it is and wait for official BIOS updates.
Cheers
NeverBackDown
AMD Enthusiast
Tom posted the 505 BIOS here.
Careful. This won't be covered by warranty. It is not officially released by Asus.
