ASRock X570S PG Riptide

[ir_cow]

The new ASRock X570S PG Riptide takes the X570 chipset to a whole new level. Gone is the controversial fan that creates unnecessary noise. The Riptide is an affordable motherboard with 2.5 Gb/s LAN from Killer Networking, RGB support, and all the necessities to get started!

Show full review

 

[tabascosauz]

New mobo reviewer? I like the expansion of the new memory OC section. Tip - use Zentimings for showing all your settings in one place, RM messy af.

Couple this with the SoC power stages not having a heatsink and memory overclocking becomes a game of chance. Without raising the SoC voltage above the default 1.1 V, it might not be possible to run DDR4-3600 to DDR4-4000 in synchronous mode with the Infinity Fabric clock. I highly suggest making sure direct airflow moves over the entire VRM section during long periods of use.

2 x 50A Vishays for SoC should still be alright? It's really only the APUs that can draw serious current through SOC, chiplet CPUs hang out in the 20-25W SOC region at most. Though yeah, ASRock cheaping out on a second heatsink is a bad move but not surprising. And recycling the Pro4 heatsink is sad.

In any case, after fiddling with it long enough using a G.Skill DDR4-5066 2x 8 GB kit, I managed to find the soft-limit of this board, which is DDR4-4800.

Can disable Geardown and try straight 2T if going for pure freq, maybe it'll go higher maybe it won't. This sounds like ASRock is still using a 4-layer PCB for this, given the price? 4800 would be pretty ass for a new 6-layer. But Steel Legend topped out at like 4666/4733 so they must have made some optimizations to the topology.

Anyhow, great review!

 

[ir_cow]

Yep I'm the new guy. I appreciate all feedback. It only makes the content better for all.

Does Zentimings lists the voltages? Thats why I like Ryzen Master. Easy to copy someone exact settings.

As for DDR4-5066, I may not not changed it to T2 and just left it on Gear Down. I'll make sure to try both next time. The voltagr drop was pissing me off that day. I could have forgotten to try it all. DDR4-4800 didnt want to post for a good while. I figured it out in the end.

 

[tabascosauz]

Yep I'm the new guy. I appreciate all feedback. It only makes the content better for all.

Does Zentimings lists the voltages? Thats why I like Ryzen Master. Easy to copy someone exact settings.

It should but it's highly CPU and board dependent so it might not give you everything. Ryzen 3000 doesn't show 2 VDDGs. APUs don't show VDDG at all. Asus boards don't show VDIMM sensor. etc

The problem with RM there is that if it's not supposed to be visible to software, I'm not sure if the voltages shown in RM are actually accurate. iirc on 3700X it felt like RM was just repeating the number I set in BIOS which as you know may change eg. for VDIMM/VSOC/VDDG.



I hear u on the vdroop. My Gigabyte ITX board has close to 0.1V of droop on SOC and LLC basically doesn't work, CPU LLC is almost as bad.

 

[ir_cow]

Oh cool! Yeah I'll give it try. I'm planning on sticking a multimeter in a few places. After this Big vdroop issue, can I really trust it reported anything accurately? The expensive boards, sure. This one...eh?

It won't make it for the B550 Riptide, I'm already past the testing phase.

 

[Dristun]

As before I feel like low-end X570 boards are kinda pointless because you can, like, just get a good B550 that's also cheaper. If someone wants many faster SSDs they will probably also opt for expansion cards that take the PCIe slot, that second m2 wired to gen4 won't make much of a difference if you really want LOTS of fast storage, IMHO.

 

[Ernest1ca]

This board does not impress me much. Needs to be dirt cheap in order to buy.

 

[W1zzard]

vdroop

Any thoughts on whether this is worth adding to our standard mobo review layout? Any ideas how to test? What about boards than don't expose any LLC setting?

 

[Makaveli]

Ya the cons on this board is far too many. Better AM4 options out there, and thanks for the review I like how you have the USB chart broken out so you can see if its coming from CPU or chipset.

 

[lukart]

This board does not impress me much. Needs to be dirt cheap in order to buy.

I think what the guy did at Asrock was creating a basic budget X570, many people are just like the review said, looking for something to plug and play, not to mess around with settings and overclocking.
Let's be real, the big part of the market don't even overclock, they just want something that works and at the right price which I believe what's this motherboard is all about.

 

[tabascosauz]

Any thoughts on whether this is worth adding to our standard mobo review layout? Any ideas how to test? What about boards than don't expose any LLC setting?

Pretty easy to test every board at the same static OC (eg 4.2@1.25v) but I feel like it would take a ridiculous amount of time to test every LLC config, usually like 6 or 7 levels. And like u said some budget locked boards have no control. Maybe only dig into it a little bit as part of the Overclocking section if the board seems to have particular issues with LLC that warrant a little snooping, like @ir_cow ran into here and has time to test?

Lott of OCers are of the opinion that more volts + relaxed LLC = better since volt regulation is usually better. So to properly test performance that would need an oscilloscope...but that's an outright insane burden to ask of any tester to keep that up for more than 1 board.

 

[Chrispy_]

6x PCIe means this is a great mining board that can also do other stuff.

 

[ir_cow]

6x PCIe means this is a great mining board that can also do other stuff.

Only if those 6 devices are not pulling 70 watts from each PCIe slot

A small update to my investigation. I did more testing as it relates to the B550 as well. Both use the same VRM design. Anyways with a Input of 1.35v (LLC 1 Highest), the back of the coil with a multimeter reads 1.430v, Software Reads SV12 1.288v. I tried a few different memory kits. Enabling XMP still reboots the computer on the 4 kits I tried. However I measured no voltage drop with the multimeter which surprised me. Software however goes down to 1.2v right before.

I also found out XMP or manually settings does not make a difference it still reboots the computer running Prime95. However, If I lower it to 2933 it does not crash. Sometimes 3200 would go for 5 minutes, but eventually it reboots. I have a direct fan on all of this for testing. No need to complicate things more with heat related issues. Something is clearly going on with voltages. Its not a hard freeze. It is a reboot after X amount of time. 5-30 seconds.

I said the reboot is due to a large vdroop I observed. I still believe this to be the case, but the software readings and the readouts conflict in the wrong way. Should I remove this comment from the review? I have ruled out this is a faulty board, that is unless the B550 Riptide is faulty as well.

 

[tabascosauz]

Only if those 6 devices are not pulling 70 watts from each PCIe slot

A small update to my investigation. I did more testing as it relates to the B550 as well. Both use the same VRM design. Anyways with a Input of 1.35v (LLC 1 Highest), the back of the coil with a multimeter reads 1.430v, Software Reads SV12 1.288v. I tried a few different memory kits. Enabling XMP still reboots the computer on the 4 kits I tried. However I measured no voltage drop with the multimeter which surprised me. Software however goes down to 1.2v right before.

I also found out XMP or manually settings does not make a difference it still reboots the computer running Prime95. However, If I lower it to 2933 it does not crash. Sometimes 3200 would go for 5 minutes, but eventually it reboots. I have a direct fan on all of this for testing. No need to complicate things more with heat related issues. Something is clearly going on with voltages. Its not a hard freeze. It is a reboot after X amount of time. 5-30 seconds.

I said the reboot is due to a large vdroop I observed. I still believe this to be the case, but the software readings and the readouts conflict in the wrong way. Should I remove this comment from the review?

It's fine to take the SVI2 readout as gospel, it's pretty accurate. The multimeter might be reading before VRM losses - this isn't a particularly high end VRM.

That said 1.35 down to 1.288 is still pretty hefty a droop for the strongest LLC, but perhaps not unexpected from a budget board with probably an unimpressive power plane and a UP9505 PWM.

Incidentally I was fighting with P95 reboots when I had a B550M Steel Legend for a short time. Almost exact same VRM with UP9505 and SiC654 doubled through UP1911s. Power plane is laid out a bit differently but parts the same my bad, it's identical. It wouldn't reboot at JEDEC speeds but anything over 3200, I suspect it was something up with VSOC or minor voltages. I didn't encounter a single other board that could run 3600 with plenty of headroom at 1.1v VSOC on my 3700X.

 

[ir_cow]

@tabascosauz I re-ran the memory test since I was already here. The reason why DDR4-5000 does work is because 2T does not take lol. I set it in the BIOS. Still 1T in windows. Gear Down works though.... So that is probably why I couldn't get 5000 to boot. Simply because the BIOS ignored my settings. Man things are looking worse for ASRock as I dig into this board more.

I need to move on though. We get it. its a budget board. No one should expect wonders from it.

 

[tabascosauz]

@tabascosauz I re-ran the memory test since I was already here. The reason why DDR4-5000 does work is because 2T does not take lol. I set it in the BIOS. Still 1T in windows. Gear Down works though.... So that is probably why I couldn't get 5000 to boot. Simply because the BIOS ignored my settings. Man things are looking worse for ASRock as I dig into this board more.

I need to move on though. We get it. its a budget board. No one should expect wonders from it.

I'm guessing you have to review with launch BIOS? Sounds like 1201 though I'm not sure, since it lists Cezanne support. That's pretty crazy, even for a budget board I've never had an AM4 board that's too bugged to accept 2T. Can't remember, is Geardown in the same dropdown as 1T and 2T on ASRock, or is it its own option?

Looked through the QVL and it looks like ASRock is lying again? Saw one 4600 kit on the Cezanne list but Vermeer/Matisse/Renoir seem to all be QVL'd only for 4266 or so.

A far cry from the DDR4-5000 displayed proudly on the product page. Hopefully they update it.

 

[ir_cow]

Its super common to see memory listed that never go that high in practice. Usually it has a star* next to it saying its CPU dependent, aka IMC. Other times its what one DIMM can do on the best of days.

I was using the newest BIOS (1.3). It could be a bug in this one. So if someone really wants 2T, well try all the different versions. I'm not going to back a 3rd time and test all the versions.

 

[Sandbo]

I really like the x1 slots have opening for longer cards, will be a life saver for me.
Too bad I have got the Gigabyte X570S UD that does not have it.

 

[TheinsanegamerN]

As before I feel like low-end X570 boards are kinda pointless because you can, like, just get a good B550 that's also cheaper. If someone wants many faster SSDs they will probably also opt for expansion cards that take the PCIe slot, that second m2 wired to gen4 won't make much of a difference if you really want LOTS of fast storage, IMHO.

Same, but I feel X570 itself is entirely unnecessary. B550 gets you PCIe 4.0 for the GPU and a single SSD. For 99% of people a single PCIe 4.0 drive is more then enough. Dual GPU is beyond dead, card makers have apparently 0 interest in making 4.0x1 10Gbps ethernet cards, internal sound cards are already 3.0x1, and nothing short of multi NVMe cards need 4.0 speeds. And I feel if you are the kind of user that needs multiple 7+GB/S drives for something like video editing then you may be more interested in a threadripper system for the extra lanes.

If B650 gets PCIe 4.0 for all attached lanes there will be 0 reasons for consumers to buy X670 other then bragging rights. As a bonus the B series micro ATX boards are actually quite good, like the MSI B550 mortar.

 

[Chrispy_]

Only if those 6 devices are not pulling 70 watts from each PCIe slot

Mining risers don't take power from the board, other than what's needed for signalling.

 

[eidairaman1]

The charts need to be organized by lowest to highest values.

Also this being an AsRock PGS board it should be compared to their top PGS board and then the Aqua/Taichi/Creator.

Mining risers don't take power from the board, other than what's needed for signalling.

View attachment 220285

All that latency through USB...

 

[ir_cow]

The charts need to be organized by lowest to highest values.

Also this being an AsRock PGS board it should be compared to their top PGS board and then the Aqua/Taichi/Creator.

The top of each chart is the best. Some charts are "lower is better" other is higher. Its easy to follow knowing that whatever is first is the "best". I wouldnt put much weight into the benchmarks. Unless something is seriously wrong, they should be all simliar...

The important stuff is; audio, vrm design, memory support, overclocking and features. I would remove comparison benchmarks altogether if I was allowed.

As for ASRock comparison, Im not following you completely on this one. Are you saying this review should only compare ASRock boards? Or do you want a dedicated page of comparing the different SKU?

Either is not possible. I am not Nate and do not have acess to past review hardware. Maybe in the distant future this could happen. But my stance is simply, go read the other review if you need to compare.