G.SKILL Flare X5 DDR5-6000 CL32 (AMD) 2x 16 GB

[ir_cow]

G.SKILL brings forward the Flare X line to the DDR5 market. In today's review we will be using the new EXPO 6000 MT/s profile that's optimized for Zen 4 using modest primary timings of 32-38-38-96. This kit is designed and marketed for AMD and targets the "sweet spot" that is getting all the attention. Follow along as we test this memory and see how it stacks up the competition!

Show full review

 

[Space Lynx]

Great review as always, keep up the good work!

 

[Psychoholic]

Exact same memory i have, they have been flawless.
I went for Flare X5 because the original Flare X worked so well with Previous Zen processors.

 

[noel_fs]

no 4k benchmarks? or are the results too meaningless?

 

[igralec84]

How many 1000% in karhu memtest is considered stable for daily use?

I'm currently experimenting with my kingston 5600 cl40 m-die running 6000 CL30-36-36-76-112 at 1.4v on top of aggressive timings preset and got an error at 2100% (1.35v got an error at 80%). Forgot to run HWinfo to check the temp but 1.4v didn't go above 50-51°C the last time i ran it for 1.5h.

Haven't touched the other voltages yet so might have some reserve without loosening timings. For the first time got over 80k read and under 65 latency

 

[firex]

Anyone tried using 4 of the 16gb ones to get 64GB total?

 

[ir_cow]

How many 1000% in karhu memtest is considered stable for daily use?

Honestly - 1000% plus a run of 2.5b y-cruncher. After that the only thing you have to worry about is DRAM temps. Because passing when you first boot is easier because its cold. Being on for 8 hours / in-use can change the outcome. It isn't the 8 hours part so much, but the fact its at normal operating temperature for a extended period of time. You don't have to use karhu memtest. I just found it errors out a lot quicker vs MemTestPro or TM5. AIDA64 and Memtest86 is useless for DDR5. Those two will "pass" and instantly fail those other 3 programs.

Anyone tried using 4 of the 16gb ones to get 64GB total?

Yes. Not with these specific ones. The outcome is the same. It depends on the motherboard, CPU and which platform your on (Intel or AMD).

no 4k benchmarks? or are the results too meaningless?

Basically. It doesn't improve memory reviews because its 100% GPU bound. When the frame rate is the same between the GPU reviews and this, I don't see a reason to include them. Not until I get a more powerful video card and it shows a real impact on perf.

 

[mechtech]

Honestly - 1000% plus a run of 2.5b y-cruncher. After that the only thing you have to worry about is DRAM temps. Because passing when you first boot is easier because its cold. Being on for 8 hours / in-use can change the outcome. It isn't the 8 hours part so much, but the fact its at normal operating temperature for a extended period of time. You don't have to use karhu memtest. I just found it errors out a lot quicker vs MemTestPro or TM5. AIDA64 and Memtest86 is useless for DDR5. Those two will "pass" and instantly fail those other 3 programs.

Good to know!! Thanks for the info. Does the same hold true for DDR4 or strictly DDR5?

Also I see memtest 86+ has been updated, same difference though?

And does thaiphoon burner support DDR5 yet?

I have to assemble and test a new PC shortly (DDR4 ryzen 5700x) and want to put it through its paces before giving it the green light, been a long time, so kind of rusty with most trusted and accurate tools/benchmarks.

Thanks

 

[ir_cow]

Good to know!! Thanks for the info. Does the same hold true for DDR4 or strictly DDR5?

AIDA64 has never really caught memory errors, but memtest86 was pretty good for DDR4. Things would slip by ever so often though. I also used it for the first round before going into windows. Now I gotten smarter with time management and have second OS drive that can be corrupted and I wont care.

Also I see memtest 86+ has been updated, same difference though?

Version 9 came built into the ASUS Z690 Hero. It didn't fair any better. Maybe a newer version is out now that is?

And does thaiphoon burner support DDR5 yet?

I haven't check recently. It was erasing / corrupting SPD data for DDR5 early on so I avoid it together now. Not everyone has access to AIDA64 :/

 

[mechtech]

Version 9 came built into the ASUS Z690 Hero. It didn't fair any better. Maybe a newer version is out now that is?

hmmm v6?
Memtest86+ v6.00
Released: 2022-10-23

Memtest86+ | The Open-Source Memory Testing Tool

Memtest86+ is an advanced, free, open-source, stand-alone memory tester for 32- and 64-bit computers (UEFI & BIOS supported)

memtest.org

10.2 on non +

MemTest86 - Download now!

Download your free copy of MemTest86 here for USB.

www.memtest86.com

 

[ir_cow]

hmmm v6?
Memtest86+ v6.00
Released: 2022-10-23

Memtest86+ | The Open-Source Memory Testing Tool

Memtest86+ is an advanced, free, open-source, stand-alone memory tester for 32- and 64-bit computers (UEFI & BIOS supported)

memtest.org

10.2 on non +

MemTest86 - Download now!

Download your free copy of MemTest86 here for USB.

www.memtest86.com

I was using v7 for DDR4 (what it says when your in the software). The Hero said 9.0. I guess V10 is out now.

Looks like a lot has been changed. https://www.memtest86.com/whats-new.html

 

[xSneak]

Would like to see the results on an esports title with competitive settings used i.e. low settings and 1% low frame rate results. I think cs go has a built in benchmark.
As you know, there are 500hz 1080p monitors coming out that are going to be cpu limited.
Thanks for the review.

 

[Chaitanya]

Yes. Not with these specific ones. The outcome is the same. It depends on the motherboard, CPU and which platform your on (Intel or AMD).

Given these are AM5 optimized SKU, do they work with all 4 slots occupied on AM5 boards? looking at QVL for most AM5 boards it seems like for now none of these CPUs are capable of running with all 4 slots fully populated.

 

[DrCR]

Any sense on pricing i.e. what a kit like this is anticipated to cost, say, 6 months or a year from now?

I’m waiting for some pricing sanity before doing an AM5 build, but I may as well build sooner rather than later if I’d have to wait too long for more better value pricing.

 

[ir_cow]

Would like to see the results on an esports title with competitive settings used i.e. low settings and 1% low frame rate results. I think cs go has a built in benchmark.
As you know, there are 500hz 1080p monitors coming out that are going to be cpu limited.
Thanks for the review.

It's been discussed. Biggest hurtle is how often esports maps change. Hard to get a repeatable test if the location you used is gone next month. Call of Duty is new every year (warzone 6 months). Fornite (every season). Valrant (already 600+ FPS), PUBG (changes randomly?), RB6S (maps change and leave rotation yearly). CS:GO might be the only one that hasn't changed much, but a empty map vs a full map has different FPS. Which map to pick?

Given these are AM5 optimized SKU, do they work with all 4 slots occupied on AM5 boards? looking at QVL for most AM5 boards it seems like for now none of these CPUs are capable of running with all 4 slots fully populated.

Officially AMD doesn't support over DDR5-3600 using 4 DIMMs. I don't have 128GB (4x32GB) to test, but it seems 4800 is probably the max. 4x16GB, it varies per MB from 5600-6200. Unless you need to capacity it is best to wait until the next Ryzen (Zen4+/Zen5).

Any sense on pricing i.e. what a kit like this is anticipated to cost, say, 6 months or a year from now?

I’m waiting for some pricing sanity before doing an AM5 build, but I may as well build sooner rather than later if I’d have to wait too long for more better value pricing.

It more reasonable now. Before DDR5 came out DDR4 64GB was $300, now DDR5 64GB is $250-350 as well. DDR4 32GB was anywhere from $150-300. DDR5 is the same. DDR4 is on a fire sale right now making it hard to compare, everyone is dumping stock. Give it another year and DDR4 will start to go up in price and be less in stock.

 

[codex5600x]

Which is the best timing foar ryzen 7 serie to use for ddr5 6000MHz?

 

[igralec84]

Which is the best timing foar ryzen 7 serie to use for ddr5 6000MHz?

CL30-36-36-74-108 or maybe CL28 from this review, if you have the suitable memory IC that can pull it off.

 

[ir_cow]

Which is the best timing foar ryzen 7 serie to use for ddr5 6000MHz?

The lowest you can go with the voltage you are willing to apply. Its a stupid answer, but there is just too many factors at play to give one answer. CPU IMC, DRAM Voltage, DRAM ICs, etc. Primary have a impact, but so does the secondary's. That is really kinda wacky on AMD right now. I haven't figured it totally out what is lowest possible yet.

Pretty much every Hynix M-Die I've tested can do 28-36-36-50-90 with either 1.45 or 1.5v for 6000 MT/s. Probably are not ever going lower then CL28 either. Maybe 34 for the the other two? Its not enough perf gain to waste time on when it will be at the edge of stability.

DDR5 scales up. I know the Flare X5 is for AMD, same principles apply here. For example (GKSILL Trident Z5 A-DIE) on Intel where I can go a lot higher. CL30 works at 6800 and CL36 for 8000 MT/s. Doesn't matter how low the freq gets, CL28 is the lowest it will go.

 

[Mussels]

Just like DDR4, it seems that getting dual rank sticks is key to performance
look how 2x32 smashed all the 2x16, despite far worse timings - and the lowest voltage of the lot

CL40 beating CL30?
The era of paying for timings is over, just get the higher capacity and call it a day

Game tests are showing something along the lines of 150 to 160FPS in memory sensitive titles, with CL30 to CL40
It exists, sure - but to such a small extent that the extra ranks are certainly overcoming a large latency difference

buying four cheap ram sticks is definitely an option

 

[AusWolf]

I had high hopes of this stick as being one of the very few that's low enough on bling that I can actually imagine it in my PC, and as also having the highly praised Hynix chips on it...

... only to see that its latencies are even looser than on my Corsair kit with the Samsung chips.

Not that it's a deal-breaker, though - just nothing to go out of one's way to buy instead of my Corsair.

Just like DDR4, it seems that getting dual rank sticks is key to performance
look how 2x32 smashed all the 2x16, despite far worse timings - and the lowest voltage of the lot

Not necessarily.

 

[mukumi]

Any sense on pricing i.e. what a kit like this is anticipated to cost, say, 6 months or a year from now?

I’m waiting for some pricing sanity before doing an AM5 build, but I may as well build sooner rather than later if I’d have to wait too long for more better value pricing.

Purchased mine 198 euros. Not excessive knowing that a good 32gb 3600c16 is around 120 euros.

 

[Mussels]

I had high hopes of this stick as being one of the very few that's low enough on bling that I can actually imagine it in my PC, and as also having the highly praised Hynix chips on it...

... only to see that its latencies are even looser than on my Corsair kit with the Samsung chips.

Not that it's a deal-breaker, though - just nothing to go out of one's way to buy instead of my Corsair.


Not necessarily.
View attachment 276940

"up to" is exactly my point, what timings are those modules running at? JEDEC? slower?
Was it tested with hynix samsung or micron?

Again, those values are "best case" and assumptions with zero guarantees of success, hence the fact that they're all labelled as overclocked because they do not work with stock settings

 

[dewd]

going to use it on intel. i accidentally bought this kit like 2 days before this review because i wasnt paying attention to expo. ive never had an amd cpu and didnt know they have xmp now. anyway, where should i start in the bios setting to make em not run on said default 4800 mhz? can i just use the values from the expo profile and manually set these timings? do i need to adjust more numbers than the main 4? not trying to beat hipro5 just wanting to get closer to what i paid for because a 4800 kit would have cost half.

z790
13700k

 

[ir_cow]

going to use it on intel. i accidentally bought this kit like 2 days before this review because i wasnt paying attention to expo. ive never had an amd cpu and didnt know they have xmp now. anyway, where should i start in the bios setting to make em not run on said default 4800 mhz? can i just use the values from the expo profile and manually set these timings? do i need to adjust more numbers than the main 4? not trying to beat hipro5 just wanting to get closer to what i paid for because a 4800 kit would have cost half.

z790
13700k

A lot of Z790 have added EXPO support now and needs a BIOS update. Can't say all of them do, but update your to the newest and see if it gained support. EXPO and XMP are basically the same thing, just the profile HEX data is moved around so that aren't directly compatible.

The profile includes the memory voltage and some secondary timings too, so if you are going to manually enter all the data, make sure to do that as well.

 

[AusWolf]

"up to" is exactly my point, what timings are those modules running at? JEDEC? slower?
Was it tested with hynix samsung or micron?

Well, considering that this G.Skill runs at 6000 MHz 32-38-38-96-130 with Hynix chips, and my Corsair is 6000 MHz 36-36-36-76-112 with Samsung, I have to say that Hynix chips don't seem to be as good as the hype around them. Maybe my Corsair is at its limit by default, and the G.Skill can be tuned further, but why bother? Like you said, the differences are minimal.

Again, those values are "best case" and assumptions with zero guarantees of success, hence the fact that they're all labelled as overclocked because they do not work with stock settings

My point is that dual-rank RAM might not necessarily result in more performance, as some boards can only do lower speeds with dual rank modules, which might have a bigger (negative) effect on performance.