runtime: reorder bits in heap bitmap bytes

The runtime deals with 1-bit pointer bitmaps and 2-bit heap bitmaps
that have entries for both pointers and mark bits.

Each byte in a 1-bit pointer bitmap looks like pppppppp (all pointer bits).
Each byte in a 2-bit heap bitmap looks like mpmpmpmp (mark, pointer, ...).
This means that when converting from 1-bit to 2-bit, as we do
during malloc, we have to pick up 4 bits in pppp form and use
shifts to create the mpmpmpmp form.

This CL changes the 2-bit heap bitmap form to mmmmpppp,
so that 4 bits picked up in 1-bit form can be used directly in
the low bits of the heap bitmap byte, without expansion.
This simplifies the code, and it also happens to be faster.

name                    old mean              new mean              delta
SetTypePtr              14.0ns × (0.98,1.09)  14.0ns × (0.98,1.08)     ~    (p=0.966)
SetTypePtr8             16.5ns × (0.99,1.05)  15.3ns × (0.96,1.16)   -6.86% (p=0.012)
SetTypePtr16            21.3ns × (0.98,1.05)  18.8ns × (0.94,1.14)  -11.49% (p=0.000)
SetTypePtr32            34.6ns × (0.93,1.22)  27.7ns × (0.91,1.26)  -20.08% (p=0.001)
SetTypePtr64            55.7ns × (0.97,1.11)  41.6ns × (0.98,1.04)  -25.30% (p=0.000)
SetTypePtr126           98.0ns × (1.00,1.00)  67.7ns × (0.99,1.05)  -30.88% (p=0.000)
SetTypePtr128           98.6ns × (1.00,1.01)  68.6ns × (0.99,1.03)  -30.44% (p=0.000)
SetTypePtrSlice          781ns × (0.99,1.01)   571ns × (0.99,1.04)  -26.93% (p=0.000)
SetTypeNode1            13.1ns × (0.99,1.01)  12.1ns × (0.99,1.01)   -7.45% (p=0.000)
SetTypeNode1Slice        113ns × (0.99,1.01)    94ns × (1.00,1.00)  -16.35% (p=0.000)
SetTypeNode8            32.7ns × (1.00,1.00)  29.8ns × (0.99,1.01)   -8.97% (p=0.000)
SetTypeNode8Slice        266ns × (1.00,1.00)   204ns × (1.00,1.00)  -23.40% (p=0.000)
SetTypeNode64           58.0ns × (0.98,1.08)  42.8ns × (1.00,1.01)  -26.24% (p=0.000)
SetTypeNode64Slice      1.55µs × (0.99,1.02)  0.96µs × (1.00,1.00)  -37.84% (p=0.000)
SetTypeNode64Dead       13.1ns × (0.99,1.01)  12.1ns × (1.00,1.00)   -7.33% (p=0.000)
SetTypeNode64DeadSlice  1.52µs × (1.00,1.01)  1.08µs × (1.00,1.01)  -28.95% (p=0.000)
SetTypeNode124          97.9ns × (1.00,1.00)  67.1ns × (1.00,1.01)  -31.49% (p=0.000)
SetTypeNode124Slice     2.87µs × (0.99,1.02)  1.75µs × (1.00,1.01)  -39.15% (p=0.000)
SetTypeNode126          98.4ns × (1.00,1.01)  68.1ns × (1.00,1.01)  -30.79% (p=0.000)
SetTypeNode126Slice     2.91µs × (0.99,1.01)  1.77µs × (0.99,1.01)  -39.09% (p=0.000)
SetTypeNode1024          732ns × (1.00,1.00)   511ns × (0.87,1.42)  -30.14% (p=0.000)
SetTypeNode1024Slice    23.1µs × (1.00,1.00)  13.9µs × (0.99,1.02)  -39.83% (p=0.000)

Change-Id: I12e3b850a4e6fa6c8146b8635ff728f3ef658819
Reviewed-on: https://go-review.googlesource.com/9828Reviewed-by: Austin Clements <austin@google.com>