I am trying to use the Intel XTU software to do some tweaking to my system.
I figured that would be easier than trying to decipher the complex BIOS settings on this Workstation Motherboard.
I am familiar with overclocking consumer boards, but the BIOS settings in the workstation board have me baffled.
Negative voltage (undervolting) is now locked in XTU due to Plundervolt vulnerability, and XTU will refuse to run without UVP enabled. Undervolting can still be done via BIOS, however. (My understanding is that W790 boards have UVP enabled by default and cannot be turned off, as you already have access to full FIVR control in the BIOS).
So I guess I would have to undervolt to get the temperature back under control.
But it does stop at 94C and start to throttle down the clocks to around 4.6GHz.
So I might try running this for a bit, I rarely do all-core workloads for extended periods.
The computer is also plugged into an APC 1500 UPS with digital display, and the UPS shows that it is drawing the same amount whether I have the Core Ratio on Auto or By Core Usage.
So the power on current draw must be a really fast transient peak that is just enough to trigger the RTX-3090 power LEDs but too quick to be shown on the UPS display.
But in By Core Usage mode, it does a two second power on with fans and RGB, then a reset as if BIOS training where the fans and RGB go off again for half a second, then back to a regular boot to BIOS screen etc.
The ASUS W790 ACE board always does the boot-reset-boot sequence when CPU By Core Usage mode is enabled.
So it must not be saving any settings that it gets from this process, and requires them on every startup.
I would say that I hope they change this in a future BIOS version but I doubt it.
So I would have to put up with the GPU Power LEDs flashing if I use this mode.
Overclocking these chips on ambient temperatures (traditional liquid or air cooling) is ridiculous. Single threaded benchmarks like Superpi32m can loop easily over 6100 Mhz without using what I consider ridiculous voltages. Multi-threaded benches like Cinebench R23 on most of the units I've tried can run at 5.7 GHz on all performance threads while set at under 1.3 volts.
You can easily speculate where the future 13900KS, which promises up to 6 GHz at stock, will end up frequency-wise especially if Intel is skimming the best dies to save for those currently. You are still going to benefit from the Thermal Grizzly OC Frame and will of course want to buy the best thermal paste that you feel is affordable to you.
What does it take to cool the Core i9-13900K? You can use an inexpensive solution if you want, but the quality of your experience on these chips is directly proportional to the quality of your cooling.
What am I trying to say? Intel sets the limits on your thermals and wants to RMA as few units as possible, so it has built in some tolerance even over 100C. So even though it feels like pushing the envelope to higher temps, the chip will be just fine running at 100C. Enjoy your high frequencies, Your Raptor Lake CPU will adapt itself voltage and clock wise to its conditions.
A CPU capable of working at -192C is the first requirement. If it is -189C or even -190C, it is not enough and is rejected. You can have a wonderful unit that does 5.8 GHz using very low voltage on ambient cooling, that then has a cold bug at -185c that is deemed useless for extreme overclocking. This is a painful process and has contributed greatly to my gray hair.
In the first three parts of this review of the Taiko Audio SGM Extreme (linked above), I looked at the Extreme out of the box, and the remarkable performance it is capable of. In this Part 4 of my review, my focus shifts to if and how to make the Extreme better by tweaking factors outside the Extreme itself (networking, power, etc).
For networking, Taiko recommends a low RFI network connection (impact will vary based on your local setup and circumstances). The Extreme comes with a StarTech optical network card, and Taiko recommends that those looking for the ultimate performance explore whether an optical network connection works better than copper ethernet for them.
Lastly, the Extreme is a beast, and requires a stable platform. Taiko recommends that it be on a stable shelf or platform as close to the floor as possible to minimize adverse impacts from mechanical vibrations.
As a reminder, my tests so far of the Extreme have been with a generic power cord plugged into a generic power strip (with all my other AV equipment and various SMPS plugged into it) which was in turn plugged into a generic outlet on a shared 15A circuit behind my media console.
The performance of the Extreme with this level of (lack of) attention to power hygiene is remarkable, and a testament to the quality of the design of the Extreme power network and casing. That being said, can we make it better by providing the Extreme better cleaner power and more power?
Going from a shared circuit to a dedicated 30A circuit, even with the generic power cord, the increase in speed and dynamics is absolutely stunning (wow!) but with perhaps a touch of edge or roughness.
When I first put in my dedicated circuits, it was dollar for dollar one of the best system upgrades I had implemented. That being said, the impact of a dedicated circuit on the Extreme is well beyond what I expected. Backgrounds are much blacker, and the physicality and control of dynamics are kicked WAY up. Details are more vivid and tangible, and depth resolution and spatial placement even more refined and extended. Everything that I was hearing before is simply better, faster, more controlled, and more coherent. Fantastic!
I have two dedicated circuits (one for my DAC and analog components, one for my digital chain) for a reason. Consolidating my entire digital chain into a single box with the Extreme alone on a dedicated circuit is an absolutely wonderful place to be. Does power at the wall matter for the Extreme? Most definitely, most emphatically yes!
As with all things power, every home situation and neighborhood situation is different. The choice between a power conditioner or power regenerator or large battery or isolation transformer or any combination of these will heavily depend on your home electrical setup, and what is going on in the electrical grid in your neighborhood. Experiment and see what works for you, but keep in mind that the better and faster and more power you deliver to the Extreme, the happier it is (think of it as a very large monoblock amplifier, not a computer).
Still going directly to a dedicated circuit outlet (no isolation transformer...my digital audio Topaz is now officially retired), next I auditioned a series of upgraded power cords that I had available to me. I just a couple days before this listening test had received one of the traveling demo Sablon Audio Prince power cords (well broken in and very well traveled exterior) as a loaner, so I was excited to hear what it could do in my system.
Going from the generic power cord to a Pangea AC-14SE, I immediately noticed a smoother presentation, with a slight(?) apparent loss in dynamics or speed (difficult to decouple some of the changes in my listening tests). There were significantly better dynamics and speed than I had heard with the Topaz isolation transformer. Certainly a very relaxed presentation, and a step forward from the generic power cord.
The Extreme uses a single Schurter 4A fuse. Given the positive impact that other power tweaks have had, is there an upside to using something like the Synergistic Research Orange fuse? As I am able to source a Synergistic Research Orange fuse to audition, I will update this review.
The Extreme uses a single Schurter 4A fuse. Given the positive impact that other power tweaks have had, is there an upside to using something like the Synergistic Research Orange fuse? After a bit of a pandemic-induced delay, I was able to source a 5A SR Orange fuse and try it with the Extreme.
At first listen, there is indeed even more detail to be had, but with perhaps a hint more of harshness? The general recommendation seems to be 100-150 hours to properly run in these fuses (the SR folks suggest as much as 200-300 hours), so time to let the stew stew a bit. After a couple days I was no longer picking up any harshness, and after a week, things were completely clean and smooth.
Again a reminder, my tests so far of the Extreme have been with a generic Cat 5 ethernet cable connected to a shared SMPS-powered generic switch in my media console, connected over a length of generic Cat 5 home ethernet to an Ubiquiti EdgeRouter 10X home router connected to an AT&T-provided optical network terminal and Pace gateway/router for my fiber internet connection.
As I noted above, in this configuration there was a notable gap between content played from local storage on the Extreme and content from on my home Mac Mini file server. There was also a much bigger step down to content streamed from TIDAL.
For the component level tests, I limited my testing to TIDAL streaming content only to force music to stream from the Internet through all the intermediate network connections and devices. When I got to an optimal configuration for streaming, I came back to compare streaming to file server content to local content with an optimized network.
When I first received my Uptone Audio EtherREGEN, I was stunned by the sound quality impact network connectivity could have. As surprised as I was by the impact the EtherREGEN was having, I was shocked at how incredibly audible differences between ethernet cables were. I have no idea how and it is a difficult conversation to have even with dedicated audiophiles, but for me, with the EtherREGEN I could tell apart ethernet cables with just a couple seconds of listening. Incredible, and an experience that forced me to reconsider a lot of things that I had taken as givens.
As an added benefit, on some of my best recordings, the low space resonances from the room where the recording was made all snapped into focus for me, giving me a very strong sense of the space where the recording was made. That feeling of being in the cathedral or the club or concert hall is awesome. The icing on the cake was that I was able to hear all this with the EtherREGEN and inexpensive but quality Cat 8 ethernet cables ($10-15 range).
c80f0f1006