- Mar 3, 2017
- 1,623
- 5,894
- 136
>non-JEDEC specNot at 6000 MT/s memory:
@DisEnchantment reported his score with 6000 MT/s memory.>non-JEDEC spec
Come on.
You're smarter than that.
we're talking mobile chips yet you shoehorn some overclocked garbage none of them will ever run.@DisEnchantment reported his score with 6000 MT/s memory.
So that is what we will use.
No, it doesn’t.
Golden Cove & Zen 4 have near identical IPC.
The rate at which you're going, Zen 5 IPC might as well depend on the a multitude of likely factors - ranging from who is likely to be the next POTUS, what happens to crude oil prices due to the middle-east situation, whether an earthquake strikes Taiwan on the day of Computex, if something more objectionable than pictures of bikini models is found with Taiwanese workers in TSMC's campus in Arizona etc.we're talking mobile chips yet you shoehorn some overclocked garbage none of them will ever run.
Like at least try to cope properly.
If you're trying to entertain me before the launch at least do it properly.The rate at which you're going, Zen 5 IPC might as well depend on the a multitude of likely factors - ranging from who is likely to be the next POTUS, what happens to crude oil prices due to the middle-east situation, whether an earthquake strikes Taiwan on the day of Computex, if something more objectionable than pictures of bikini models is found with Taiwanese workers in TSMC's campus in Arizona etc.
Very, GB5/6 hate singlechannels.I would like to note that the lower IPC score has 8GB of ram. Probably this means that the memory interface is not fully populated. I have no idea how this impacts geekbench scores.
So the first number is IPC, second is clocks, and third is IPC x Clocks. This puts Z5 at 15% ahead of Lion Cove in ST, which is decent but I was hoping for a bigger smack down, like closer to 20%, due to the rumors of ARL being 10% faster than RPL and Z5 being 30% faster than Z4.
Perhaps @adroc_thurston was actually sandbagging Z5 improvement🤣
He's not modeling it correctly.So the first number is IPC, second is clocks, and third is IPC x Clocks. This puts Z5 at 15% ahead of Lion Cove in ST, which is decent but I was hoping for a bigger smack down, like closer to 20%, due to the rumors of ARL being 10% faster than RPL and Z5 being 30% faster than Z4.
In which direction? As you mentioned, IPC does not scale linearly with clocks so one can argue the Z5 projections are higher than they should be.He's not modeling it correctly.
Wrong baseline.In which direction? As you mentioned, IPC does not scale linearly with clocks so one can argue the Z5 projections are higher than they should be.
It doesn't make a difference, I get same ST score with one stick and two sticks (MT score was me doing something when I see ST tests done)I would like to note that the lower IPC score has 8GB of ram. Probably this means that the memory interface is not fully populated. I have no idea how this impacts geekbench scores.
Well, the geekbench browser is full of hvm results that all seem to detect frequencies correctly, so it should.I will fire up Xen or QEMU if GB can really detect properly cpu frequency properly, which I doubt but would verify if I have time.
But maybe not worthwhile since it is just an ES, stepping 0 at that, and probably microcode would be restrictive for outside distribution
In this data there is like a double digit IPC difference between ZEN4 Mobile (1.28) , ZEN4 Desktop(1.36) and ZEN4 X3D (1.45). So the data seems to be completely wrong overall. Probably just took some random result of everything and compared it.So the first number is IPC, second is clocks, and third is IPC x Clocks. This puts Z5 at 15% ahead of Lion Cove in ST, which is decent but I was hoping for a bigger smack down, like closer to 20%, due to the rumors of ARL being 10% faster than RPL and Z5 being 30% faster than Z4.
In which direction? As you mentioned, IPC does not scale linearly with clocks so one can argue the Z5 projections are higher than they should be.
It does, they just don't idle at all and the v/f is dingus.Doesn't mean it did it for an early engineering sample though
I did it, lscpu shows 2.6 Base for my 7950X, but cpufreq reports 2G, I don't have GB account, so maybe someone can check the frequencies
My RAM is 6000MT/s
*snip*
That's a very good point. Good catch.In this data there is like a double digit IPC difference between ZEN4 Mobile (1.28) , ZEN4 Desktop(1.36) and ZEN4 X3D (1.45). So the data seems to be completely wrong overall. Probably just took some random result of everything and compared it.
3050 6 GB scores around 4800 graphics score in Time Spy.Will Strix Point iGPU be able match/exceed RTX 3050 mobile?
This might be a hot-take. In my opinion, a 128 bit APU should be able to match atleast a 2 generation old RTX xx50 dGPU.
Ugh, no.If Strix Point would have that 16 MB IC as was rumored before by some individuals - it would touch 5000 pts.
Why does that matter at all when every sample was tested with the same RAM?>non-JEDEC spec
Come on.
You're smarter than that.